use crate::agent_log::AgentLogWriter;
use crate::slog;
use crate::slog_error;
use crate::slog_warn;
use crate::config::{AgentConfig, ProjectConfig};
use crate::worktree::{self, WorktreeInfo};
use portable_pty::{ChildKiller, CommandBuilder, PtySize, native_pty_system};
use serde::Serialize;
use std::collections::HashMap;
use std::io::{BufRead, BufReader};
use std::path::{Path, PathBuf};
use std::process::Command;
use std::sync::{Arc, Mutex};
use std::time::Instant;
use tokio::sync::broadcast;

/// Default TTL for completed/failed agent entries: 1 hour.
pub const DEFAULT_AGENT_TTL_SECS: u64 = 3600;

/// Events emitted during server startup reconciliation to broadcast real-time
/// progress to connected WebSocket clients.
#[derive(Debug, Clone, Serialize)]
pub struct ReconciliationEvent {
    /// The story being reconciled, or empty string for the overall "done" event.
    pub story_id: String,
    /// Coarse status: "checking", "gates_running", "advanced", "skipped", "failed", "done"
    pub status: String,
    /// Human-readable details.
    pub message: String,
}

/// Build the composite key used to track agents in the pool.
fn composite_key(story_id: &str, agent_name: &str) -> String {
    format!("{story_id}:{agent_name}")
}

/// RAII guard that removes a pending agent entry from the pool on drop.
///
/// Created after inserting a `Pending` entry into the agent HashMap.
/// If `start_agent` succeeds (the agent process is spawned and status
/// transitions to `Running`), call [`disarm`](Self::disarm) to prevent
/// cleanup.  If any intermediate step fails and the guard is dropped
/// without being disarmed, the pending entry is removed so it cannot
/// block future auto-assign dispatches.
struct PendingGuard {
    agents: Arc<Mutex<HashMap<String, StoryAgent>>>,
    key: String,
    armed: bool,
}

impl PendingGuard {
    fn new(agents: Arc<Mutex<HashMap<String, StoryAgent>>>, key: String) -> Self {
        Self {
            agents,
            key,
            armed: true,
        }
    }

    /// Prevent the guard from cleaning up the entry (call after
    /// successful spawn).
    fn disarm(&mut self) {
        self.armed = false;
    }
}

impl Drop for PendingGuard {
    fn drop(&mut self) {
        if self.armed
            && let Ok(mut agents) = self.agents.lock()
            && agents
                .get(&self.key)
                .is_some_and(|a| a.status == AgentStatus::Pending)
        {
            agents.remove(&self.key);
            slog!(
                "[agents] Cleaned up leaked Pending entry for '{}'",
                self.key
            );
        }
    }
}

/// Events streamed from a running agent to SSE clients.
#[derive(Debug, Clone, Serialize)]
#[serde(tag = "type", rename_all = "snake_case")]
pub enum AgentEvent {
    /// Agent status changed.
    Status {
        story_id: String,
        agent_name: String,
        status: String,
    },
    /// Raw text output from the agent process.
    Output {
        story_id: String,
        agent_name: String,
        text: String,
    },
    /// Agent produced a JSON event from `--output-format stream-json`.
    AgentJson {
        story_id: String,
        agent_name: String,
        data: serde_json::Value,
    },
    /// Agent finished.
    Done {
        story_id: String,
        agent_name: String,
        session_id: Option<String>,
    },
    /// Agent errored.
    Error {
        story_id: String,
        agent_name: String,
        message: String,
    },
    /// Thinking tokens from an extended-thinking block.
    Thinking {
        story_id: String,
        agent_name: String,
        text: String,
    },
}

#[derive(Debug, Clone, Serialize, PartialEq)]
#[serde(rename_all = "snake_case")]
pub enum AgentStatus {
    Pending,
    Running,
    Completed,
    Failed,
}

impl std::fmt::Display for AgentStatus {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        match self {
            Self::Pending => write!(f, "pending"),
            Self::Running => write!(f, "running"),
            Self::Completed => write!(f, "completed"),
            Self::Failed => write!(f, "failed"),
        }
    }
}

/// Pipeline stages for automatic story advancement.
#[derive(Debug, Clone, PartialEq)]
pub enum PipelineStage {
    /// Coding agents (coder-1, coder-2, etc.)
    Coder,
    /// QA review agent
    Qa,
    /// Mergemaster agent
    Mergemaster,
    /// Supervisors and unknown agents — no automatic advancement.
    Other,
}

/// Determine the pipeline stage from an agent name.
pub fn pipeline_stage(agent_name: &str) -> PipelineStage {
    match agent_name {
        "qa" => PipelineStage::Qa,
        "mergemaster" => PipelineStage::Mergemaster,
        name if name.starts_with("coder") => PipelineStage::Coder,
        _ => PipelineStage::Other,
    }
}

/// Determine the pipeline stage for a configured agent.
///
/// Prefers the explicit `stage` config field (added in Bug 150) over the
/// legacy name-based heuristic so that agents with non-standard names
/// (e.g. `qa-2`, `coder-opus`) are assigned to the correct stage.
fn agent_config_stage(cfg: &AgentConfig) -> PipelineStage {
    match cfg.stage.as_deref() {
        Some("coder") => PipelineStage::Coder,
        Some("qa") => PipelineStage::Qa,
        Some("mergemaster") => PipelineStage::Mergemaster,
        Some(_) => PipelineStage::Other,
        None => pipeline_stage(&cfg.name),
    }
}

/// Completion report produced when acceptance gates are run.
///
/// Created automatically by the server when an agent process exits normally,
/// or via the internal `report_completion` method.
#[derive(Debug, Serialize, Clone)]
pub struct CompletionReport {
    pub summary: String,
    pub gates_passed: bool,
    pub gate_output: String,
}

#[derive(Debug, Serialize, Clone)]
pub struct AgentInfo {
    pub story_id: String,
    pub agent_name: String,
    pub status: AgentStatus,
    pub session_id: Option<String>,
    pub worktree_path: Option<String>,
    pub base_branch: Option<String>,
    pub completion: Option<CompletionReport>,
    /// UUID identifying the persistent log file for this session.
    pub log_session_id: Option<String>,
}

struct StoryAgent {
    agent_name: String,
    status: AgentStatus,
    worktree_info: Option<WorktreeInfo>,
    session_id: Option<String>,
    tx: broadcast::Sender<AgentEvent>,
    task_handle: Option<tokio::task::JoinHandle<()>>,
    /// Accumulated events for polling via get_agent_output.
    event_log: Arc<Mutex<Vec<AgentEvent>>>,
    /// Set when the agent calls report_completion.
    completion: Option<CompletionReport>,
    /// Project root, stored for pipeline advancement after completion.
    project_root: Option<PathBuf>,
    /// UUID identifying the log file for this session.
    log_session_id: Option<String>,
    /// Timestamp when the agent entered a terminal state (Completed/Failed).
    /// Used by the TTL reaper to remove stale entries.
    completed_at: Option<Instant>,
}

/// Build an `AgentInfo` snapshot from a `StoryAgent` map entry.
fn agent_info_from_entry(story_id: &str, agent: &StoryAgent) -> AgentInfo {
    AgentInfo {
        story_id: story_id.to_string(),
        agent_name: agent.agent_name.clone(),
        status: agent.status.clone(),
        session_id: agent.session_id.clone(),
        worktree_path: agent
            .worktree_info
            .as_ref()
            .map(|wt| wt.path.to_string_lossy().to_string()),
        base_branch: agent
            .worktree_info
            .as_ref()
            .map(|wt| wt.base_branch.clone()),
        completion: agent.completion.clone(),
        log_session_id: agent.log_session_id.clone(),
    }
}

/// Manages concurrent story agents, each in its own worktree.
pub struct AgentPool {
    agents: Arc<Mutex<HashMap<String, StoryAgent>>>,
    port: u16,
    /// Registry of active PTY child process killers, keyed by "{story_id}:{agent_name}".
    /// Used to terminate child processes on server shutdown or agent stop, preventing
    /// orphaned Claude Code processes from running after the server exits.
    child_killers: Arc<Mutex<HashMap<String, Box<dyn ChildKiller + Send + Sync>>>>,
}

/// RAII guard that removes a child killer from the registry on drop.
///
/// This ensures the killer is always cleaned up when `run_agent_pty_blocking`
/// returns, regardless of the exit path (normal completion, timeout, or error).
struct ChildKillerGuard {
    killers: Arc<Mutex<HashMap<String, Box<dyn ChildKiller + Send + Sync>>>>,
    key: String,
}

impl Drop for ChildKillerGuard {
    fn drop(&mut self) {
        if let Ok(mut killers) = self.killers.lock() {
            killers.remove(&self.key);
        }
    }
}

impl AgentPool {
    pub fn new(port: u16) -> Self {
        Self {
            agents: Arc::new(Mutex::new(HashMap::new())),
            port,
            child_killers: Arc::new(Mutex::new(HashMap::new())),
        }
    }

    /// Kill all active PTY child processes.
    ///
    /// Called on server shutdown to prevent orphaned Claude Code processes from
    /// continuing to run after the server exits. Each registered killer is called
    /// once, then the registry is cleared.
    pub fn kill_all_children(&self) {
        if let Ok(mut killers) = self.child_killers.lock() {
            for (key, killer) in killers.iter_mut() {
                slog!("[agents] Killing child process for {key} on shutdown");
                let _ = killer.kill();
            }
            killers.clear();
        }
    }

    /// Kill and deregister the child process for a specific agent key.
    ///
    /// Used by `stop_agent` to ensure the PTY child is terminated even though
    /// aborting a `spawn_blocking` task handle does not interrupt the blocking thread.
    fn kill_child_for_key(&self, key: &str) {
        if let Ok(mut killers) = self.child_killers.lock()
            && let Some(mut killer) = killers.remove(key)
        {
            slog!("[agents] Killing child process for {key} on stop");
            let _ = killer.kill();
        }
    }

    /// Start an agent for a story: load config, create worktree, spawn agent.
    /// If `agent_name` is None, defaults to the first configured agent.
    /// If `resume_context` is provided, it is appended to the rendered prompt
    /// so the agent can pick up from a previous failed attempt.
    pub async fn start_agent(
        &self,
        project_root: &Path,
        story_id: &str,
        agent_name: Option<&str>,
        resume_context: Option<&str>,
    ) -> Result<AgentInfo, String> {
        let config = ProjectConfig::load(project_root)?;

        // Resolve agent name from config
        let resolved_name = match agent_name {
            Some(name) => {
                config
                    .find_agent(name)
                    .ok_or_else(|| format!("No agent named '{name}' in config"))?;
                name.to_string()
            }
            None => config
                .default_agent()
                .ok_or_else(|| "No agents configured".to_string())?
                .name
                .clone(),
        };

        let key = composite_key(story_id, &resolved_name);

        // Create shared resources before the atomic check-and-insert so the
        // agents lock is held continuously from the availability check through
        // the Pending insert, eliminating the TOCTOU race (story 132).
        let (tx, _) = broadcast::channel::<AgentEvent>(1024);

        let event_log: Arc<Mutex<Vec<AgentEvent>>> = Arc::new(Mutex::new(Vec::new()));

        // Generate a unique session ID for the persistent log file.
        let log_session_id = uuid::Uuid::new_v4().to_string();

        // Create persistent log writer.
        let log_writer = match AgentLogWriter::new(
            project_root,
            story_id,
            &resolved_name,
            &log_session_id,
        ) {
            Ok(w) => Some(Arc::new(Mutex::new(w))),
            Err(e) => {
                eprintln!("[agents] Failed to create log writer for {story_id}:{resolved_name}: {e}");
                None
            }
        };

        // Atomically check availability and register as Pending.  The lock is
        // held continuously from the duplicate check through the HashMap insert
        // so no concurrent start_agent call can slip through the check before
        // this insert completes (fixes TOCTOU race, story 132).
        //
        // The `PendingGuard` ensures that if any step below fails the entry is
        // removed from the pool so it does not permanently block auto-assign
        // (bug 118).
        {
            let mut agents = self.agents.lock().map_err(|e| e.to_string())?;
            if let Some(agent) = agents.get(&key)
                && (agent.status == AgentStatus::Running || agent.status == AgentStatus::Pending)
            {
                return Err(format!(
                    "Agent '{resolved_name}' for story '{story_id}' is already {}",
                    agent.status
                ));
            }
            // Enforce single-instance concurrency: if this agent is already running on
            // any other story, reject the request. The story remains in its current pipeline
            // directory and `auto_assign_available_work` will pick it up when the agent
            // becomes free.
            if let Some(busy_story) = agents.iter().find_map(|(k, a)| {
                if a.agent_name == resolved_name
                    && matches!(a.status, AgentStatus::Running | AgentStatus::Pending)
                {
                    Some(
                        k.rsplit_once(':')
                            .map(|(sid, _)| sid)
                            .unwrap_or(k)
                            .to_string(),
                    )
                } else {
                    None
                }
            }) {
                return Err(format!(
                    "Agent '{resolved_name}' is already running on story '{busy_story}'; \
                     story '{story_id}' will be picked up when the agent becomes available"
                ));
            }
            agents.insert(
                key.clone(),
                StoryAgent {
                    agent_name: resolved_name.clone(),
                    status: AgentStatus::Pending,
                    worktree_info: None,
                    session_id: None,
                    tx: tx.clone(),
                    task_handle: None,
                    event_log: event_log.clone(),
                    completion: None,
                    project_root: Some(project_root.to_path_buf()),
                    log_session_id: Some(log_session_id.clone()),
                    completed_at: None,
                },
            );
        }
        let mut pending_guard = PendingGuard::new(self.agents.clone(), key.clone());

        let _ = tx.send(AgentEvent::Status {
            story_id: story_id.to_string(),
            agent_name: resolved_name.clone(),
            status: "pending".to_string(),
        });

        // Move story from upcoming/ to current/ and auto-commit before creating the worktree.
        move_story_to_current(project_root, story_id)?;

        // Extract inactivity timeout from the agent config before cloning config.
        let inactivity_timeout_secs = config
            .find_agent(&resolved_name)
            .map(|a| a.inactivity_timeout_secs)
            .unwrap_or(300);

        // Clone all values needed inside the background spawn.
        let project_root_clone = project_root.to_path_buf();
        let config_clone = config.clone();
        let resume_context_owned = resume_context.map(str::to_string);
        let sid = story_id.to_string();
        let aname = resolved_name.clone();
        let tx_clone = tx.clone();
        let agents_ref = self.agents.clone();
        let key_clone = key.clone();
        let log_clone = event_log.clone();
        let port_for_task = self.port;
        let log_writer_clone = log_writer.clone();
        let child_killers_clone = self.child_killers.clone();

        // Spawn the background task. Worktree creation and agent launch happen here
        // so `start_agent` returns immediately after registering the agent as
        // Pending — non-blocking by design (story 157).
        let handle = tokio::spawn(async move {
            // Step 1: create the worktree (slow — git checkout, pnpm install, etc.)
            let wt_info = match worktree::create_worktree(
                &project_root_clone,
                &sid,
                &config_clone,
                port_for_task,
            )
            .await
            {
                Ok(wt) => wt,
                Err(e) => {
                    // Worktree creation failed — mark agent as Failed so the UI shows the error.
                    if let Ok(mut agents) = agents_ref.lock()
                        && let Some(agent) = agents.get_mut(&key_clone)
                    {
                        agent.status = AgentStatus::Failed;
                        agent.completed_at = Some(Instant::now());
                    }
                    let _ = tx_clone.send(AgentEvent::Error {
                        story_id: sid.clone(),
                        agent_name: aname.clone(),
                        message: format!("Failed to create worktree: {e}"),
                    });
                    return;
                }
            };

            // Step 2: store worktree info and render agent command/args/prompt.
            let wt_path_str = wt_info.path.to_string_lossy().to_string();
            {
                if let Ok(mut agents) = agents_ref.lock()
                    && let Some(agent) = agents.get_mut(&key_clone)
                {
                    agent.worktree_info = Some(wt_info.clone());
                }
            }

            let (command, args, mut prompt) = match config_clone.render_agent_args(
                &wt_path_str,
                &sid,
                Some(&aname),
                Some(&wt_info.base_branch),
            ) {
                Ok(result) => result,
                Err(e) => {
                    if let Ok(mut agents) = agents_ref.lock()
                        && let Some(agent) = agents.get_mut(&key_clone)
                    {
                        agent.status = AgentStatus::Failed;
                        agent.completed_at = Some(Instant::now());
                    }
                    let _ = tx_clone.send(AgentEvent::Error {
                        story_id: sid.clone(),
                        agent_name: aname.clone(),
                        message: format!("Failed to render agent args: {e}"),
                    });
                    return;
                }
            };

            // Append resume context if this is a restart with failure information.
            if let Some(ctx) = resume_context_owned {
                prompt.push_str(&ctx);
            }

            // Step 3: transition to Running now that the worktree is ready.
            {
                if let Ok(mut agents) = agents_ref.lock()
                    && let Some(agent) = agents.get_mut(&key_clone)
                {
                    agent.status = AgentStatus::Running;
                }
            }
            let _ = tx_clone.send(AgentEvent::Status {
                story_id: sid.clone(),
                agent_name: aname.clone(),
                status: "running".to_string(),
            });

            // Step 4: launch the agent process.
            match run_agent_pty_streaming(
                &sid,
                &aname,
                &command,
                &args,
                &prompt,
                &wt_path_str,
                &tx_clone,
                &log_clone,
                log_writer_clone,
                inactivity_timeout_secs,
                child_killers_clone,
            )
            .await
            {
                Ok(session_id) => {
                    // Server-owned completion: run acceptance gates automatically
                    // when the agent process exits normally.
                    run_server_owned_completion(
                        &agents_ref,
                        port_for_task,
                        &sid,
                        &aname,
                        session_id,
                    )
                    .await;
                }
                Err(e) => {
                    if let Ok(mut agents) = agents_ref.lock()
                        && let Some(agent) = agents.get_mut(&key_clone)
                    {
                        agent.status = AgentStatus::Failed;
                        agent.completed_at = Some(Instant::now());
                    }
                    let _ = tx_clone.send(AgentEvent::Error {
                        story_id: sid.clone(),
                        agent_name: aname.clone(),
                        message: e,
                    });
                }
            }
        });

        // Store the task handle while the agent is still Pending.
        {
            let mut agents = self.agents.lock().map_err(|e| e.to_string())?;
            if let Some(agent) = agents.get_mut(&key) {
                agent.task_handle = Some(handle);
            }
        }

        // Agent successfully spawned — prevent the guard from removing the entry.
        pending_guard.disarm();

        Ok(AgentInfo {
            story_id: story_id.to_string(),
            agent_name: resolved_name,
            status: AgentStatus::Pending,
            session_id: None,
            worktree_path: None,
            base_branch: None,
            completion: None,
            log_session_id: Some(log_session_id),
        })
    }

    /// Stop a running agent. Worktree is preserved for inspection.
    pub async fn stop_agent(
        &self,
        _project_root: &Path,
        story_id: &str,
        agent_name: &str,
    ) -> Result<(), String> {
        let key = composite_key(story_id, agent_name);

        let (worktree_info, task_handle, tx) = {
            let mut agents = self.agents.lock().map_err(|e| e.to_string())?;
            let agent = agents
                .get_mut(&key)
                .ok_or_else(|| format!("No agent '{agent_name}' for story '{story_id}'"))?;

            let wt = agent.worktree_info.clone();
            let handle = agent.task_handle.take();
            let tx = agent.tx.clone();
            agent.status = AgentStatus::Failed;
            (wt, handle, tx)
        };

        // Abort the task and kill the PTY child process.
        // Note: aborting a spawn_blocking task handle does not interrupt the blocking
        // thread, so we must also kill the child process directly via the killer registry.
        if let Some(handle) = task_handle {
            handle.abort();
            let _ = handle.await;
        }
        self.kill_child_for_key(&key);

        // Preserve worktree for inspection — don't destroy agent's work on stop.
        if let Some(ref wt) = worktree_info {
            slog!(
                "[agents] Worktree preserved for {story_id}:{agent_name}: {}",
                wt.path.display()
            );
        }

        let _ = tx.send(AgentEvent::Status {
            story_id: story_id.to_string(),
            agent_name: agent_name.to_string(),
            status: "stopped".to_string(),
        });

        // Remove from map
        {
            let mut agents = self.agents.lock().map_err(|e| e.to_string())?;
            agents.remove(&key);
        }

        Ok(())
    }

    /// List all agents with their status.
    pub fn list_agents(&self) -> Result<Vec<AgentInfo>, String> {
        let agents = self.agents.lock().map_err(|e| e.to_string())?;
        Ok(agents
            .iter()
            .map(|(key, agent)| {
                // Extract story_id from composite key "story_id:agent_name"
                let story_id = key
                    .rsplit_once(':')
                    .map(|(sid, _)| sid.to_string())
                    .unwrap_or_else(|| key.clone());
                agent_info_from_entry(&story_id, agent)
            })
            .collect())
    }

    /// Subscribe to events for a story agent.
    pub fn subscribe(
        &self,
        story_id: &str,
        agent_name: &str,
    ) -> Result<broadcast::Receiver<AgentEvent>, String> {
        let key = composite_key(story_id, agent_name);
        let agents = self.agents.lock().map_err(|e| e.to_string())?;
        let agent = agents
            .get(&key)
            .ok_or_else(|| format!("No agent '{agent_name}' for story '{story_id}'"))?;
        Ok(agent.tx.subscribe())
    }

    /// Drain accumulated events for polling. Returns all events since the last drain.
    pub fn drain_events(
        &self,
        story_id: &str,
        agent_name: &str,
    ) -> Result<Vec<AgentEvent>, String> {
        let key = composite_key(story_id, agent_name);
        let agents = self.agents.lock().map_err(|e| e.to_string())?;
        let agent = agents
            .get(&key)
            .ok_or_else(|| format!("No agent '{agent_name}' for story '{story_id}'"))?;
        let mut log = agent.event_log.lock().map_err(|e| e.to_string())?;
        Ok(log.drain(..).collect())
    }

    /// Block until the agent reaches a terminal state (completed, failed, stopped).
    /// Returns the agent's final `AgentInfo`.
    /// `timeout_ms` caps how long to wait; returns an error if the deadline passes.
    pub async fn wait_for_agent(
        &self,
        story_id: &str,
        agent_name: &str,
        timeout_ms: u64,
    ) -> Result<AgentInfo, String> {
        // Subscribe before checking status so we don't miss the terminal event
        // if the agent completes in the window between the two operations.
        let mut rx = self.subscribe(story_id, agent_name)?;

        // Return immediately if already in a terminal state.
        {
            let agents = self.agents.lock().map_err(|e| e.to_string())?;
            let key = composite_key(story_id, agent_name);
            if let Some(agent) = agents.get(&key)
                && matches!(agent.status, AgentStatus::Completed | AgentStatus::Failed)
            {
                return Ok(agent_info_from_entry(story_id, agent));
            }
        }

        let deadline =
            tokio::time::Instant::now() + std::time::Duration::from_millis(timeout_ms);

        loop {
            let remaining = deadline.saturating_duration_since(tokio::time::Instant::now());
            if remaining.is_zero() {
                return Err(format!(
                    "Timed out after {timeout_ms}ms waiting for agent '{agent_name}' on story '{story_id}'"
                ));
            }

            match tokio::time::timeout(remaining, rx.recv()).await {
                Ok(Ok(event)) => {
                    let is_terminal = match &event {
                        AgentEvent::Done { .. } | AgentEvent::Error { .. } => true,
                        AgentEvent::Status { status, .. } if status == "stopped" => true,
                        _ => false,
                    };
                    if is_terminal {
                        let agents = self.agents.lock().map_err(|e| e.to_string())?;
                        let key = composite_key(story_id, agent_name);
                        return Ok(if let Some(agent) = agents.get(&key) {
                            agent_info_from_entry(story_id, agent)
                        } else {
                            // Agent was removed from map (e.g. stop_agent removes it after
                            // the "stopped" status event is sent).
                            let (status, session_id) = match event {
                                AgentEvent::Done { session_id, .. } => {
                                    (AgentStatus::Completed, session_id)
                                }
                                _ => (AgentStatus::Failed, None),
                            };
                            AgentInfo {
                                story_id: story_id.to_string(),
                                agent_name: agent_name.to_string(),
                                status,
                                session_id,
                                worktree_path: None,
                                base_branch: None,
                                completion: None,
                                log_session_id: None,
                            }
                        });
                    }
                }
                Ok(Err(broadcast::error::RecvError::Lagged(_))) => {
                    // Missed some buffered events — check current status before resuming.
                    let agents = self.agents.lock().map_err(|e| e.to_string())?;
                    let key = composite_key(story_id, agent_name);
                    if let Some(agent) = agents.get(&key)
                        && matches!(agent.status, AgentStatus::Completed | AgentStatus::Failed)
                    {
                        return Ok(agent_info_from_entry(story_id, agent));
                    }
                    // Still running — continue the loop.
                }
                Ok(Err(broadcast::error::RecvError::Closed)) => {
                    // Channel closed: no more events will arrive. Return current state.
                    let agents = self.agents.lock().map_err(|e| e.to_string())?;
                    let key = composite_key(story_id, agent_name);
                    if let Some(agent) = agents.get(&key) {
                        return Ok(agent_info_from_entry(story_id, agent));
                    }
                    return Err(format!(
                        "Agent '{agent_name}' for story '{story_id}' channel closed unexpectedly"
                    ));
                }
                Err(_) => {
                    return Err(format!(
                        "Timed out after {timeout_ms}ms waiting for agent '{agent_name}' on story '{story_id}'"
                    ));
                }
            }
        }
    }

    /// Create a worktree for the given story using the server port (writes .mcp.json).
    pub async fn create_worktree(
        &self,
        project_root: &Path,
        story_id: &str,
    ) -> Result<worktree::WorktreeInfo, String> {
        let config = ProjectConfig::load(project_root)?;
        worktree::create_worktree(project_root, story_id, &config, self.port).await
    }

    /// Advance the pipeline after an agent completes.
    ///
    /// Called internally by `report_completion` as a background task.
    /// Reads the stored completion report and project_root from the agent,
    /// then drives the next pipeline stage based on the agent's role:
    ///
    /// - **Coder** + gates passed → move story to `work/3_qa/`, start `qa` agent.
    /// - **Coder** + gates failed → restart the same coder agent with failure context.
    /// - **QA** + gates passed + coverage passed → move story to `work/4_merge/`, start `mergemaster` agent.
    /// - **QA** + gates passed + coverage failed → restart `qa` with coverage failure context.
    /// - **QA** + gates failed → restart `qa` with failure context.
    /// - **Mergemaster** → run `script/test` on master; if pass: archive + cleanup worktree;
    ///   if fail: restart `mergemaster` with failure context.
    /// - **Other** (supervisor, unknown) → no automatic advancement.
    async fn run_pipeline_advance_for_completed_agent(&self, story_id: &str, agent_name: &str) {
        let key = composite_key(story_id, agent_name);

        let (completion, project_root, worktree_path) = {
            let agents = match self.agents.lock() {
                Ok(a) => a,
                Err(e) => {
                    slog_error!("[pipeline] Failed to lock agents for '{story_id}:{agent_name}': {e}");
                    return;
                }
            };
            let agent = match agents.get(&key) {
                Some(a) => a,
                None => return,
            };
            let wt_path = agent
                .worktree_info
                .as_ref()
                .map(|wt| wt.path.clone());
            (agent.completion.clone(), agent.project_root.clone(), wt_path)
        };

        let completion = match completion {
            Some(c) => c,
            None => {
                slog_warn!("[pipeline] No completion report for '{story_id}:{agent_name}'");
                return;
            }
        };
        let project_root = match project_root {
            Some(p) => p,
            None => {
                slog_warn!("[pipeline] No project_root for '{story_id}:{agent_name}'");
                return;
            }
        };

        let config = ProjectConfig::load(&project_root).unwrap_or_default();
        let stage = config
            .find_agent(agent_name)
            .map(agent_config_stage)
            .unwrap_or_else(|| pipeline_stage(agent_name));

        match stage {
            PipelineStage::Other => {
                // Supervisors and unknown agents do not advance the pipeline.
            }
            PipelineStage::Coder => {
                if completion.gates_passed {
                    slog!(
                        "[pipeline] Coder '{agent_name}' passed gates for '{story_id}'. Moving to QA."
                    );
                    if let Err(e) = move_story_to_qa(&project_root, story_id) {
                        slog_error!("[pipeline] Failed to move '{story_id}' to 3_qa/: {e}");
                        return;
                    }
                    if let Err(e) = self
                        .start_agent(&project_root, story_id, Some("qa"), None)
                        .await
                    {
                        slog_error!("[pipeline] Failed to start qa agent for '{story_id}': {e}");
                    }
                    // Coder slot is now free — pick up any other unassigned work in 2_current/.
                    self.auto_assign_available_work(&project_root).await;
                } else {
                    slog!(
                        "[pipeline] Coder '{agent_name}' failed gates for '{story_id}'. Restarting."
                    );
                    let context = format!(
                        "\n\n---\n## Previous Attempt Failed\n\
                         The acceptance gates failed with the following output:\n{}\n\n\
                         Please review the failures above, fix the issues, and try again.",
                        completion.gate_output
                    );
                    if let Err(e) = self
                        .start_agent(&project_root, story_id, Some(agent_name), Some(&context))
                        .await
                    {
                        slog_error!(
                            "[pipeline] Failed to restart coder '{agent_name}' for '{story_id}': {e}"
                        );
                    }
                }
            }
            PipelineStage::Qa => {
                if completion.gates_passed {
                    // Run coverage gate in the QA worktree before advancing to merge.
                    let coverage_path = worktree_path.clone().unwrap_or_else(|| project_root.clone());
                    let cp = coverage_path.clone();
                    let coverage_result =
                        tokio::task::spawn_blocking(move || run_coverage_gate(&cp))
                            .await
                            .unwrap_or_else(|e| {
                                slog_warn!("[pipeline] Coverage gate task panicked: {e}");
                                Ok((false, format!("Coverage gate task panicked: {e}")))
                            });
                    let (coverage_passed, coverage_output) = match coverage_result {
                        Ok(pair) => pair,
                        Err(e) => (false, e),
                    };

                    if coverage_passed {
                        slog!(
                            "[pipeline] QA passed gates and coverage for '{story_id}'. Moving to merge."
                        );
                        if let Err(e) = move_story_to_merge(&project_root, story_id) {
                            slog_error!("[pipeline] Failed to move '{story_id}' to 4_merge/: {e}");
                            return;
                        }
                        if let Err(e) = self
                            .start_agent(&project_root, story_id, Some("mergemaster"), None)
                            .await
                        {
                            slog_error!("[pipeline] Failed to start mergemaster for '{story_id}': {e}");
                        }
                        // QA slot is now free — pick up any other unassigned work in 3_qa/.
                        self.auto_assign_available_work(&project_root).await;
                    } else {
                        slog!(
                            "[pipeline] QA coverage gate failed for '{story_id}'. Restarting QA."
                        );
                        let context = format!(
                            "\n\n---\n## Coverage Gate Failed\n\
                             The coverage gate (script/test_coverage) failed with the following output:\n{}\n\n\
                             Please improve test coverage until the coverage gate passes.",
                            coverage_output
                        );
                        if let Err(e) = self
                            .start_agent(&project_root, story_id, Some("qa"), Some(&context))
                            .await
                        {
                            slog_error!("[pipeline] Failed to restart qa for '{story_id}': {e}");
                        }
                    }
                } else {
                    slog!(
                        "[pipeline] QA failed gates for '{story_id}'. Restarting."
                    );
                    let context = format!(
                        "\n\n---\n## Previous QA Attempt Failed\n\
                         The acceptance gates failed with the following output:\n{}\n\n\
                         Please re-run and fix the issues.",
                        completion.gate_output
                    );
                    if let Err(e) = self
                        .start_agent(&project_root, story_id, Some("qa"), Some(&context))
                        .await
                    {
                        slog_error!("[pipeline] Failed to restart qa for '{story_id}': {e}");
                    }
                }
            }
            PipelineStage::Mergemaster => {
                // Run script/test on master (project_root) as the post-merge verification.
                slog!(
                    "[pipeline] Mergemaster completed for '{story_id}'. Running post-merge tests on master."
                );
                let root = project_root.clone();
                let test_result = tokio::task::spawn_blocking(move || run_project_tests(&root))
                    .await
                    .unwrap_or_else(|e| {
                        slog_warn!("[pipeline] Post-merge test task panicked: {e}");
                        Ok((false, format!("Test task panicked: {e}")))
                    });
                let (passed, output) = match test_result {
                    Ok(pair) => pair,
                    Err(e) => (false, e),
                };

                if passed {
                    slog!(
                        "[pipeline] Post-merge tests passed for '{story_id}'. Moving to done."
                    );
                    if let Err(e) = move_story_to_archived(&project_root, story_id) {
                        slog_error!("[pipeline] Failed to move '{story_id}' to done: {e}");
                    }
                    self.remove_agents_for_story(story_id);
                    // Mergemaster slot is now free — pick up any other items in 4_merge/.
                    self.auto_assign_available_work(&project_root).await;
                    // TODO: Re-enable worktree cleanup once we have persistent agent logs.
                    // Removing worktrees destroys evidence needed to debug empty-commit agents.
                    // let config =
                    //     crate::config::ProjectConfig::load(&project_root).unwrap_or_default();
                    // if let Err(e) =
                    //     worktree::remove_worktree_by_story_id(&project_root, story_id, &config)
                    //         .await
                    // {
                    //     slog!(
                    //         "[pipeline] Failed to remove worktree for '{story_id}': {e}"
                    //     );
                    // }
                    slog!(
                        "[pipeline] Story '{story_id}' done. Worktree preserved for inspection."
                    );
                } else {
                    slog!(
                        "[pipeline] Post-merge tests failed for '{story_id}'. Restarting mergemaster."
                    );
                    let context = format!(
                        "\n\n---\n## Post-Merge Test Failed\n\
                         The tests on master failed with the following output:\n{}\n\n\
                         Please investigate and resolve the failures, then call merge_agent_work again.",
                        output
                    );
                    if let Err(e) = self
                        .start_agent(&project_root, story_id, Some("mergemaster"), Some(&context))
                        .await
                    {
                        slog_error!(
                            "[pipeline] Failed to restart mergemaster for '{story_id}': {e}"
                        );
                    }
                }
            }
        }
    }

    /// Internal: report that an agent has finished work on a story.
    ///
    /// **Note:** This is no longer exposed as an MCP tool. The server now
    /// automatically runs completion gates when an agent process exits
    /// (see `run_server_owned_completion`). This method is retained for
    /// backwards compatibility and testing.
    ///
    /// - Rejects with an error if the worktree has uncommitted changes.
    /// - Runs acceptance gates (cargo clippy + cargo nextest run / cargo test).
    /// - Stores the `CompletionReport` on the agent record.
    /// - Transitions status to `Completed` (gates passed) or `Failed` (gates failed).
    /// - Emits a `Done` event so `wait_for_agent` unblocks.
    #[allow(dead_code)]
    pub async fn report_completion(
        &self,
        story_id: &str,
        agent_name: &str,
        summary: &str,
    ) -> Result<CompletionReport, String> {
        let key = composite_key(story_id, agent_name);

        // Verify agent exists, is Running, and grab its worktree path.
        let worktree_path = {
            let agents = self.agents.lock().map_err(|e| e.to_string())?;
            let agent = agents
                .get(&key)
                .ok_or_else(|| format!("No agent '{agent_name}' for story '{story_id}'"))?;

            if agent.status != AgentStatus::Running {
                return Err(format!(
                    "Agent '{agent_name}' for story '{story_id}' is not running (status: {}). \
                     report_completion can only be called by a running agent.",
                    agent.status
                ));
            }

            agent
                .worktree_info
                .as_ref()
                .map(|wt| wt.path.clone())
                .ok_or_else(|| {
                    format!(
                        "Agent '{agent_name}' for story '{story_id}' has no worktree. \
                         Cannot run acceptance gates."
                    )
                })?
        };

        let path = worktree_path.clone();

        // Run gate checks in a blocking thread to avoid stalling the async runtime.
        let (gates_passed, gate_output) = tokio::task::spawn_blocking(move || {
            // Step 1: Reject if worktree is dirty.
            check_uncommitted_changes(&path)?;
            // Step 2: Run clippy + tests and return (passed, output).
            run_acceptance_gates(&path)
        })
        .await
        .map_err(|e| format!("Gate check task panicked: {e}"))??;

        let report = CompletionReport {
            summary: summary.to_string(),
            gates_passed,
            gate_output,
        };

        // Store the completion report and advance status.
        let (tx, session_id) = {
            let mut agents = self.agents.lock().map_err(|e| e.to_string())?;
            let agent = agents.get_mut(&key).ok_or_else(|| {
                format!("Agent '{agent_name}' for story '{story_id}' disappeared during gate check")
            })?;
            agent.completion = Some(report.clone());
            agent.status = if gates_passed {
                AgentStatus::Completed
            } else {
                AgentStatus::Failed
            };
            agent.completed_at = Some(Instant::now());
            (agent.tx.clone(), agent.session_id.clone())
        };

        // Emit Done so wait_for_agent unblocks.
        let _ = tx.send(AgentEvent::Done {
            story_id: story_id.to_string(),
            agent_name: agent_name.to_string(),
            session_id,
        });

        // Advance the pipeline state machine in a background task.
        // Only advance when the agent completed (not failed) to avoid spurious restarts
        // from agents that never ran acceptance gates properly.
        let pool_clone = Self {
            agents: Arc::clone(&self.agents),
            port: self.port,
            child_killers: Arc::clone(&self.child_killers),
        };
        let sid = story_id.to_string();
        let aname = agent_name.to_string();
        tokio::spawn(async move {
            pool_clone
                .run_pipeline_advance_for_completed_agent(&sid, &aname)
                .await;
        });

        Ok(report)
    }

    /// Run the full mergemaster pipeline for a completed story:
    ///
    /// 1. Squash-merge the story's feature branch into the current branch (master).
    /// 2. If conflicts are found: abort the merge and report them.
    /// 3. Quality gates run **inside the merge worktree** before master is touched.
    /// 4. If gates pass: fast-forward master and archive the story.
    ///
    /// Returns a `MergeReport` with full details of what happened.
    pub async fn merge_agent_work(
        &self,
        project_root: &Path,
        story_id: &str,
    ) -> Result<MergeReport, String> {
        let branch = format!("feature/story-{story_id}");
        let wt_path = worktree::worktree_path(project_root, story_id);
        let root = project_root.to_path_buf();
        let sid = story_id.to_string();
        let br = branch.clone();

        // Run blocking operations (git + cargo + quality gates) off the async runtime.
        // Quality gates now run inside run_squash_merge before the fast-forward.
        let merge_result =
            tokio::task::spawn_blocking(move || run_squash_merge(&root, &br, &sid))
                .await
                .map_err(|e| format!("Merge task panicked: {e}"))??;

        if !merge_result.success {
            return Ok(MergeReport {
                story_id: story_id.to_string(),
                success: false,
                had_conflicts: merge_result.had_conflicts,
                conflicts_resolved: merge_result.conflicts_resolved,
                conflict_details: merge_result.conflict_details,
                gates_passed: merge_result.gates_passed,
                gate_output: merge_result.output,
                worktree_cleaned_up: false,
                story_archived: false,
            });
        }

        // Merge + gates both passed — archive the story and clean up agent entries.
        let story_archived = move_story_to_archived(project_root, story_id).is_ok();
        if story_archived {
            self.remove_agents_for_story(story_id);
        }

        // Clean up the worktree if it exists.
        let worktree_cleaned_up = if wt_path.exists() {
            let config = crate::config::ProjectConfig::load(project_root)
                .unwrap_or_default();
            worktree::remove_worktree_by_story_id(project_root, story_id, &config)
                .await
                .is_ok()
        } else {
            false
        };

        Ok(MergeReport {
            story_id: story_id.to_string(),
            success: true,
            had_conflicts: merge_result.had_conflicts,
            conflicts_resolved: merge_result.conflicts_resolved,
            conflict_details: merge_result.conflict_details,
            gates_passed: true,
            gate_output: merge_result.output,
            worktree_cleaned_up,
            story_archived,
        })
    }

    /// Return the port this server is running on.
    #[allow(dead_code)]
    pub fn port(&self) -> u16 {
        self.port
    }

    /// Get project root helper.
    pub fn get_project_root(
        &self,
        state: &crate::state::SessionState,
    ) -> Result<PathBuf, String> {
        state.get_project_root()
    }

    /// Get the log session ID and project root for an agent, if available.
    ///
    /// Used by MCP tools to find the persistent log file for a completed agent.
    pub fn get_log_info(
        &self,
        story_id: &str,
        agent_name: &str,
    ) -> Option<(String, PathBuf)> {
        let key = composite_key(story_id, agent_name);
        let agents = self.agents.lock().ok()?;
        let agent = agents.get(&key)?;
        let session_id = agent.log_session_id.clone()?;
        let project_root = agent.project_root.clone()?;
        Some((session_id, project_root))
    }

    /// Test helper: inject a pre-built agent entry so unit tests can exercise
    /// wait/subscribe logic without spawning a real process.
    #[cfg(test)]
    pub fn inject_test_agent(
        &self,
        story_id: &str,
        agent_name: &str,
        status: AgentStatus,
    ) -> broadcast::Sender<AgentEvent> {
        let completed_at = if matches!(status, AgentStatus::Completed | AgentStatus::Failed) {
            Some(Instant::now())
        } else {
            None
        };
        let (tx, _) = broadcast::channel::<AgentEvent>(64);
        let key = composite_key(story_id, agent_name);
        let mut agents = self.agents.lock().unwrap();
        agents.insert(
            key,
            StoryAgent {
                agent_name: agent_name.to_string(),
                status,
                worktree_info: None,
                session_id: None,
                tx: tx.clone(),
                task_handle: None,
                event_log: Arc::new(Mutex::new(Vec::new())),
                completion: None,
                project_root: None,
                log_session_id: None,
                completed_at,
            },
        );
        tx
    }

    /// Test helper: inject an agent with a specific worktree path for testing
    /// gate-related logic.
    #[cfg(test)]
    pub fn inject_test_agent_with_path(
        &self,
        story_id: &str,
        agent_name: &str,
        status: AgentStatus,
        worktree_path: PathBuf,
    ) -> broadcast::Sender<AgentEvent> {
        let completed_at = if matches!(status, AgentStatus::Completed | AgentStatus::Failed) {
            Some(Instant::now())
        } else {
            None
        };
        let (tx, _) = broadcast::channel::<AgentEvent>(64);
        let key = composite_key(story_id, agent_name);
        let mut agents = self.agents.lock().unwrap();
        agents.insert(
            key,
            StoryAgent {
                agent_name: agent_name.to_string(),
                status,
                worktree_info: Some(WorktreeInfo {
                    path: worktree_path,
                    branch: format!("feature/story-{story_id}"),
                    base_branch: "master".to_string(),
                }),
                session_id: None,
                tx: tx.clone(),
                task_handle: None,
                event_log: Arc::new(Mutex::new(Vec::new())),
                completion: None,
                project_root: None,
                log_session_id: None,
                completed_at,
            },
        );
        tx
    }

    /// Automatically assign free agents to stories waiting in the active pipeline stages.
    ///
    /// Scans `work/2_current/`, `work/3_qa/`, and `work/4_merge/` for items that have no
    /// active agent and assigns the first free agent of the appropriate role. Items in
    /// `work/1_upcoming/` are never auto-started.
    ///
    /// Respects the configured agent roster: the maximum number of concurrently active agents
    /// per role is bounded by the count of agents of that role defined in `project.toml`.
    pub async fn auto_assign_available_work(&self, project_root: &Path) {
        let config = match ProjectConfig::load(project_root) {
            Ok(c) => c,
            Err(e) => {
                slog_warn!("[auto-assign] Failed to load project config: {e}");
                return;
            }
        };

        // Process each active pipeline stage in order.
        let stages: [(&str, PipelineStage); 3] = [
            ("2_current", PipelineStage::Coder),
            ("3_qa", PipelineStage::Qa),
            ("4_merge", PipelineStage::Mergemaster),
        ];

        for (stage_dir, stage) in &stages {
            let items = scan_stage_items(project_root, stage_dir);
            if items.is_empty() {
                continue;
            }

            for story_id in &items {
                // Re-acquire the lock on each iteration to see state changes
                // from previous start_agent calls in the same pass.
                let (already_assigned, free_agent) = {
                    let agents = match self.agents.lock() {
                        Ok(a) => a,
                        Err(e) => {
                            slog_error!("[auto-assign] Failed to lock agents: {e}");
                            break;
                        }
                    };
                    let assigned = is_story_assigned_for_stage(&config, &agents, story_id, stage);
                    let free = if assigned {
                        None
                    } else {
                        find_free_agent_for_stage(&config, &agents, stage)
                            .map(|s| s.to_string())
                    };
                    (assigned, free)
                };

                if already_assigned {
                    // Story already has an active agent — skip silently.
                    continue;
                }

                match free_agent {
                    Some(agent_name) => {
                        slog!(
                            "[auto-assign] Assigning '{agent_name}' to '{story_id}' in {stage_dir}/"
                        );
                        if let Err(e) = self
                            .start_agent(project_root, story_id, Some(&agent_name), None)
                            .await
                        {
                            slog!(
                                "[auto-assign] Failed to start '{agent_name}' for '{story_id}': {e}"
                            );
                        }
                    }
                    None => {
                        // No free agents of this type — stop scanning this stage.
                        slog!(
                            "[auto-assign] All {:?} agents busy; remaining items in {stage_dir}/ will wait.",
                            stage
                        );
                        break;
                    }
                }
            }
        }
    }

    /// Reconcile stories whose agent work was committed while the server was offline.
    ///
    /// On server startup the in-memory agent pool is empty, so any story that an agent
    /// completed during a previous session is stuck: the worktree has committed work but
    /// the pipeline never advanced.  This method detects those stories, re-runs the
    /// acceptance gates, and advances the pipeline stage so that `auto_assign_available_work`
    /// (called immediately after) picks up the right next-stage agents.
    ///
    /// Algorithm:
    /// 1. List all worktree directories under `{project_root}/.story_kit/worktrees/`.
    /// 2. For each worktree, check whether its feature branch has commits ahead of the
    ///    base branch (`master` / `main`).
    /// 3. If committed work is found AND the story is in `2_current/` or `3_qa/`:
    ///    - Run acceptance gates (uncommitted-change check + clippy + tests).
    ///    - On pass + `2_current/`: move the story to `3_qa/`.
    ///    - On pass + `3_qa/`: run the coverage gate; if that also passes move to `4_merge/`.
    ///    - On failure: leave the story where it is so `auto_assign_available_work` can
    ///      start a fresh agent to retry.
    /// 4. Stories in `4_merge/` are left for `auto_assign_available_work` to handle via a
    ///    fresh mergemaster (squash-merge must be re-executed by the mergemaster agent).
    pub async fn reconcile_on_startup(
        &self,
        project_root: &Path,
        progress_tx: &broadcast::Sender<ReconciliationEvent>,
    ) {
        let worktrees = match worktree::list_worktrees(project_root) {
            Ok(wt) => wt,
            Err(e) => {
                eprintln!("[startup:reconcile] Failed to list worktrees: {e}");
                let _ = progress_tx.send(ReconciliationEvent {
                    story_id: String::new(),
                    status: "done".to_string(),
                    message: format!("Reconciliation failed: {e}"),
                });
                return;
            }
        };

        for wt_entry in &worktrees {
            let story_id = &wt_entry.story_id;
            let wt_path = wt_entry.path.clone();

            // Determine which active stage the story is in.
            let stage_dir = match find_active_story_stage(project_root, story_id) {
                Some(s) => s,
                None => continue, // Not in any active stage (upcoming/archived or unknown).
            };

            // 4_merge/ is left for auto_assign to handle with a fresh mergemaster.
            if stage_dir == "4_merge" {
                continue;
            }

            let _ = progress_tx.send(ReconciliationEvent {
                story_id: story_id.clone(),
                status: "checking".to_string(),
                message: format!("Checking for committed work in {stage_dir}/"),
            });

            // Check whether the worktree has commits ahead of the base branch.
            let wt_path_for_check = wt_path.clone();
            let has_work = tokio::task::spawn_blocking(move || {
                worktree_has_committed_work(&wt_path_for_check)
            })
            .await
            .unwrap_or(false);

            if !has_work {
                eprintln!(
                    "[startup:reconcile] No committed work for '{story_id}' in {stage_dir}/; skipping."
                );
                let _ = progress_tx.send(ReconciliationEvent {
                    story_id: story_id.clone(),
                    status: "skipped".to_string(),
                    message: "No committed work found; skipping.".to_string(),
                });
                continue;
            }

            eprintln!(
                "[startup:reconcile] Found committed work for '{story_id}' in {stage_dir}/. Running acceptance gates."
            );
            let _ = progress_tx.send(ReconciliationEvent {
                story_id: story_id.clone(),
                status: "gates_running".to_string(),
                message: "Running acceptance gates…".to_string(),
            });

            // Run acceptance gates on the worktree.
            let wt_path_for_gates = wt_path.clone();
            let gates_result = tokio::task::spawn_blocking(move || {
                check_uncommitted_changes(&wt_path_for_gates)?;
                run_acceptance_gates(&wt_path_for_gates)
            })
            .await;

            let (gates_passed, gate_output) = match gates_result {
                Ok(Ok(pair)) => pair,
                Ok(Err(e)) => {
                    eprintln!("[startup:reconcile] Gate check error for '{story_id}': {e}");
                    let _ = progress_tx.send(ReconciliationEvent {
                        story_id: story_id.clone(),
                        status: "failed".to_string(),
                        message: format!("Gate error: {e}"),
                    });
                    continue;
                }
                Err(e) => {
                    eprintln!(
                        "[startup:reconcile] Gate check task panicked for '{story_id}': {e}"
                    );
                    let _ = progress_tx.send(ReconciliationEvent {
                        story_id: story_id.clone(),
                        status: "failed".to_string(),
                        message: format!("Gate task panicked: {e}"),
                    });
                    continue;
                }
            };

            if !gates_passed {
                eprintln!(
                    "[startup:reconcile] Gates failed for '{story_id}': {gate_output}\n\
                     Leaving in {stage_dir}/ for auto-assign to restart the agent."
                );
                let _ = progress_tx.send(ReconciliationEvent {
                    story_id: story_id.clone(),
                    status: "failed".to_string(),
                    message: "Gates failed; will be retried by auto-assign.".to_string(),
                });
                continue;
            }

            eprintln!(
                "[startup:reconcile] Gates passed for '{story_id}' (stage: {stage_dir}/)."
            );

            if stage_dir == "2_current" {
                // Coder stage → advance to QA.
                if let Err(e) = move_story_to_qa(project_root, story_id) {
                    eprintln!("[startup:reconcile] Failed to move '{story_id}' to 3_qa/: {e}");
                    let _ = progress_tx.send(ReconciliationEvent {
                        story_id: story_id.clone(),
                        status: "failed".to_string(),
                        message: format!("Failed to advance to QA: {e}"),
                    });
                } else {
                    eprintln!("[startup:reconcile] Moved '{story_id}' → 3_qa/.");
                    let _ = progress_tx.send(ReconciliationEvent {
                        story_id: story_id.clone(),
                        status: "advanced".to_string(),
                        message: "Gates passed — moved to QA.".to_string(),
                    });
                }
            } else if stage_dir == "3_qa" {
                // QA stage → run coverage gate before advancing to merge.
                let wt_path_for_cov = wt_path.clone();
                let coverage_result =
                    tokio::task::spawn_blocking(move || run_coverage_gate(&wt_path_for_cov))
                        .await;

                let (coverage_passed, coverage_output) = match coverage_result {
                    Ok(Ok(pair)) => pair,
                    Ok(Err(e)) => {
                        eprintln!(
                            "[startup:reconcile] Coverage gate error for '{story_id}': {e}"
                        );
                        let _ = progress_tx.send(ReconciliationEvent {
                            story_id: story_id.clone(),
                            status: "failed".to_string(),
                            message: format!("Coverage gate error: {e}"),
                        });
                        continue;
                    }
                    Err(e) => {
                        eprintln!(
                            "[startup:reconcile] Coverage gate panicked for '{story_id}': {e}"
                        );
                        let _ = progress_tx.send(ReconciliationEvent {
                            story_id: story_id.clone(),
                            status: "failed".to_string(),
                            message: format!("Coverage gate panicked: {e}"),
                        });
                        continue;
                    }
                };

                if coverage_passed {
                    if let Err(e) = move_story_to_merge(project_root, story_id) {
                        eprintln!(
                            "[startup:reconcile] Failed to move '{story_id}' to 4_merge/: {e}"
                        );
                        let _ = progress_tx.send(ReconciliationEvent {
                            story_id: story_id.clone(),
                            status: "failed".to_string(),
                            message: format!("Failed to advance to merge: {e}"),
                        });
                    } else {
                        eprintln!("[startup:reconcile] Moved '{story_id}' → 4_merge/.");
                        let _ = progress_tx.send(ReconciliationEvent {
                            story_id: story_id.clone(),
                            status: "advanced".to_string(),
                            message: "Gates passed — moved to merge.".to_string(),
                        });
                    }
                } else {
                    eprintln!(
                        "[startup:reconcile] Coverage gate failed for '{story_id}': {coverage_output}\n\
                         Leaving in 3_qa/ for auto-assign to restart the QA agent."
                    );
                    let _ = progress_tx.send(ReconciliationEvent {
                        story_id: story_id.clone(),
                        status: "failed".to_string(),
                        message: "Coverage gate failed; will be retried.".to_string(),
                    });
                }
            }
        }

        // Signal that reconciliation is complete.
        let _ = progress_tx.send(ReconciliationEvent {
            story_id: String::new(),
            status: "done".to_string(),
            message: "Startup reconciliation complete.".to_string(),
        });
    }

    /// Test helper: inject an agent with a completion report and project_root
    /// for testing pipeline advance logic without spawning real agents.
    #[cfg(test)]
    pub fn inject_test_agent_with_completion(
        &self,
        story_id: &str,
        agent_name: &str,
        status: AgentStatus,
        project_root: PathBuf,
        completion: CompletionReport,
    ) -> broadcast::Sender<AgentEvent> {
        let completed_at = if matches!(status, AgentStatus::Completed | AgentStatus::Failed) {
            Some(Instant::now())
        } else {
            None
        };
        let (tx, _) = broadcast::channel::<AgentEvent>(64);
        let key = composite_key(story_id, agent_name);
        let mut agents = self.agents.lock().unwrap();
        agents.insert(
            key,
            StoryAgent {
                agent_name: agent_name.to_string(),
                status,
                worktree_info: None,
                session_id: None,
                tx: tx.clone(),
                task_handle: None,
                event_log: Arc::new(Mutex::new(Vec::new())),
                completion: Some(completion),
                project_root: Some(project_root),
                log_session_id: None,
                completed_at,
            },
        );
        tx
    }

    /// Inject a Running agent with a pre-built (possibly finished) task handle.
    /// Used by watchdog tests to simulate an orphaned agent.
    #[cfg(test)]
    pub fn inject_test_agent_with_handle(
        &self,
        story_id: &str,
        agent_name: &str,
        status: AgentStatus,
        task_handle: tokio::task::JoinHandle<()>,
    ) -> broadcast::Sender<AgentEvent> {
        let (tx, _) = broadcast::channel::<AgentEvent>(64);
        let key = composite_key(story_id, agent_name);
        let mut agents = self.agents.lock().unwrap();
        agents.insert(
            key,
            StoryAgent {
                agent_name: agent_name.to_string(),
                status,
                worktree_info: None,
                session_id: None,
                tx: tx.clone(),
                task_handle: Some(task_handle),
                event_log: Arc::new(Mutex::new(Vec::new())),
                completion: None,
                project_root: None,
                log_session_id: None,
                completed_at: None,
            },
        );
        tx
    }

    /// Test helper: inject a child killer into the registry.
    #[cfg(test)]
    pub fn inject_child_killer(&self, key: &str, killer: Box<dyn ChildKiller + Send + Sync>) {
        let mut killers = self.child_killers.lock().unwrap();
        killers.insert(key.to_string(), killer);
    }

    /// Test helper: return the number of registered child killers.
    #[cfg(test)]
    pub fn child_killer_count(&self) -> usize {
        self.child_killers.lock().unwrap().len()
    }

    /// Run a single watchdog pass synchronously (test helper).
    #[cfg(test)]
    pub fn run_watchdog_once(&self) {
        check_orphaned_agents(&self.agents);
    }

    /// Spawn a background watchdog task that periodically checks for Running agents
    /// whose underlying task has already finished (orphaned entries).  Any such agent
    /// is marked Failed and an Error event is emitted so that `wait_for_agent` unblocks.
    ///
    /// The watchdog runs every 30 seconds.  It is a safety net for edge cases where the
    /// PTY read loop exits without updating the agent status (e.g. a panic in the
    /// spawn_blocking task, or an external SIGKILL that closes the PTY fd immediately).
    ///
    /// When orphaned agents are detected and a `project_root` is provided, auto-assign
    /// is triggered so that free agents can pick up unassigned work.
    pub fn spawn_watchdog(pool: Arc<AgentPool>, project_root: Option<PathBuf>) {
        tokio::spawn(async move {
            let mut interval =
                tokio::time::interval(std::time::Duration::from_secs(30));
            loop {
                interval.tick().await;
                let found = check_orphaned_agents(&pool.agents);
                if found > 0
                    && let Some(ref root) = project_root
                {
                    slog!(
                        "[watchdog] {found} orphaned agent(s) detected; triggering auto-assign."
                    );
                    pool.auto_assign_available_work(root).await;
                }
            }
        });
    }

    /// Remove all agent entries for a given story_id from the pool.
    ///
    /// Called when a story is archived so that stale entries don't accumulate.
    /// Returns the number of entries removed.
    pub fn remove_agents_for_story(&self, story_id: &str) -> usize {
        let mut agents = match self.agents.lock() {
            Ok(a) => a,
            Err(e) => {
                slog_error!("[agents] Failed to lock pool for cleanup of '{story_id}': {e}");
                return 0;
            }
        };
        let prefix = format!("{story_id}:");
        let keys_to_remove: Vec<String> = agents
            .keys()
            .filter(|k| k.starts_with(&prefix))
            .cloned()
            .collect();
        let count = keys_to_remove.len();
        for key in &keys_to_remove {
            agents.remove(key);
        }
        if count > 0 {
            slog!("[agents] Removed {count} agent entries for archived story '{story_id}'");
        }
        count
    }

    /// Reap agent entries in terminal states (Completed/Failed) whose `completed_at`
    /// timestamp is older than `ttl`. Returns the number of entries reaped.
    pub fn reap_expired_agents(&self, ttl: std::time::Duration) -> usize {
        let mut agents = match self.agents.lock() {
            Ok(a) => a,
            Err(e) => {
                slog_warn!("[reaper] Failed to lock pool for TTL reaping: {e}");
                return 0;
            }
        };
        let now = Instant::now();
        let keys_to_remove: Vec<String> = agents
            .iter()
            .filter(|(_, agent)| {
                matches!(agent.status, AgentStatus::Completed | AgentStatus::Failed)
                    && agent
                        .completed_at
                        .is_some_and(|t| now.duration_since(t) >= ttl)
            })
            .map(|(k, _)| k.clone())
            .collect();
        let count = keys_to_remove.len();
        for key in &keys_to_remove {
            agents.remove(key);
        }
        if count > 0 {
            slog!("[reaper] Reaped {count} expired agent entries (TTL: {}s)", ttl.as_secs());
        }
        count
    }
}

/// Return the active pipeline stage directory name for `story_id`, or `None` if the
/// story is not in any active stage (`2_current/`, `3_qa/`, `4_merge/`).
fn find_active_story_stage(project_root: &Path, story_id: &str) -> Option<&'static str> {
    const STAGES: [&str; 3] = ["2_current", "3_qa", "4_merge"];
    for stage in &STAGES {
        let path = project_root
            .join(".story_kit")
            .join("work")
            .join(stage)
            .join(format!("{story_id}.md"));
        if path.exists() {
            return Some(stage);
        }
    }
    None
}

/// Scan a work pipeline stage directory and return story IDs, sorted alphabetically.
/// Returns an empty `Vec` if the directory does not exist.
fn scan_stage_items(project_root: &Path, stage_dir: &str) -> Vec<String> {
    let dir = project_root
        .join(".story_kit")
        .join("work")
        .join(stage_dir);
    if !dir.is_dir() {
        return Vec::new();
    }
    let mut items = Vec::new();
    if let Ok(entries) = std::fs::read_dir(&dir) {
        for entry in entries.flatten() {
            let path = entry.path();
            if path.extension().and_then(|e| e.to_str()) == Some("md")
                && let Some(stem) = path.file_stem().and_then(|s| s.to_str())
            {
                items.push(stem.to_string());
            }
        }
    }
    items.sort();
    items
}

/// Return `true` if `story_id` has any active (pending/running) agent matching `stage`.
///
/// Uses the explicit `stage` config field when the agent is found in `config`;
/// falls back to the legacy name-based heuristic for unlisted agents.
fn is_story_assigned_for_stage(
    config: &ProjectConfig,
    agents: &HashMap<String, StoryAgent>,
    story_id: &str,
    stage: &PipelineStage,
) -> bool {
    agents.iter().any(|(key, agent)| {
        // Composite key format: "{story_id}:{agent_name}"
        let key_story_id = key.rsplit_once(':').map(|(sid, _)| sid).unwrap_or(key);
        let agent_stage = config
            .find_agent(&agent.agent_name)
            .map(agent_config_stage)
            .unwrap_or_else(|| pipeline_stage(&agent.agent_name));
        key_story_id == story_id
            && agent_stage == *stage
            && matches!(agent.status, AgentStatus::Running | AgentStatus::Pending)
    })
}

/// Find the first configured agent for `stage` that has no active (pending/running) assignment.
/// Returns `None` if all agents for that stage are busy or none are configured.
/// Uses the agent's explicit `stage` config field (preferred) or falls back to name-based detection.
fn find_free_agent_for_stage<'a>(
    config: &'a ProjectConfig,
    agents: &HashMap<String, StoryAgent>,
    stage: &PipelineStage,
) -> Option<&'a str> {
    for agent_config in &config.agent {
        if agent_config_stage(agent_config) != *stage {
            continue;
        }
        let is_busy = agents.values().any(|a| {
            a.agent_name == agent_config.name
                && matches!(a.status, AgentStatus::Running | AgentStatus::Pending)
        });
        if !is_busy {
            return Some(&agent_config.name);
        }
    }
    None
}

/// Scan the agent pool for Running entries whose backing tokio task has already
/// finished and mark them as Failed.
///
/// This handles the case where the PTY read loop or the spawned task exits
/// without updating the agent status — for example when the process is killed
/// externally and the PTY master fd returns EOF before our inactivity timeout
/// fires, but some other edge case prevents the normal cleanup path from running.
fn check_orphaned_agents(agents: &Mutex<HashMap<String, StoryAgent>>) -> usize {
    let mut lock = match agents.lock() {
        Ok(l) => l,
        Err(_) => return 0,
    };

    // Collect orphaned entries: Running or Pending agents whose task handle is finished.
    // Pending agents can be orphaned if worktree creation panics before setting status.
    let orphaned: Vec<(String, String, broadcast::Sender<AgentEvent>, AgentStatus)> = lock
        .iter()
        .filter_map(|(key, agent)| {
            if matches!(agent.status, AgentStatus::Running | AgentStatus::Pending)
                && let Some(handle) = &agent.task_handle
                && handle.is_finished()
            {
                let story_id = key
                    .rsplit_once(':')
                    .map(|(s, _)| s.to_string())
                    .unwrap_or_else(|| key.clone());
                return Some((key.clone(), story_id, agent.tx.clone(), agent.status.clone()));
            }
            None
        })
        .collect();

    let count = orphaned.len();
    for (key, story_id, tx, prev_status) in orphaned {
        if let Some(agent) = lock.get_mut(&key) {
            agent.status = AgentStatus::Failed;
            slog!(
                "[watchdog] Orphaned agent '{key}': task finished but status was {prev_status}. \
                 Marking Failed."
            );
            let _ = tx.send(AgentEvent::Error {
                story_id,
                agent_name: agent.agent_name.clone(),
                message: "Agent process terminated unexpectedly (watchdog detected orphan)"
                    .to_string(),
            });
        }
    }
    count
}

/// Server-owned completion: runs acceptance gates when an agent process exits
/// normally, and advances the pipeline based on results.
///
/// This is a **free function** (not a method on `AgentPool`) to break the
/// opaque type cycle that would otherwise arise: `start_agent` → spawned task
/// → server-owned completion → pipeline advance → `start_agent`.
///
/// If the agent already has a completion report (e.g. from a legacy
/// `report_completion` call), this is a no-op to avoid double-running gates.
async fn run_server_owned_completion(
    agents: &Arc<Mutex<HashMap<String, StoryAgent>>>,
    port: u16,
    story_id: &str,
    agent_name: &str,
    session_id: Option<String>,
) {
    let key = composite_key(story_id, agent_name);

    // Guard: skip if completion was already recorded (legacy path).
    {
        let lock = match agents.lock() {
            Ok(a) => a,
            Err(_) => return,
        };
        match lock.get(&key) {
            Some(agent) if agent.completion.is_some() => {
                slog!(
                    "[agents] Completion already recorded for '{story_id}:{agent_name}'; \
                     skipping server-owned gates."
                );
                return;
            }
            Some(_) => {}
            None => return,
        }
    }

    // Get worktree path for running gates.
    let worktree_path = {
        let lock = match agents.lock() {
            Ok(a) => a,
            Err(_) => return,
        };
        lock.get(&key)
            .and_then(|a| a.worktree_info.as_ref().map(|wt| wt.path.clone()))
    };

    // Run acceptance gates.
    let (gates_passed, gate_output) = if let Some(wt_path) = worktree_path {
        let path = wt_path;
        match tokio::task::spawn_blocking(move || {
            check_uncommitted_changes(&path)?;
            run_acceptance_gates(&path)
        })
        .await
        {
            Ok(Ok(result)) => result,
            Ok(Err(e)) => (false, e),
            Err(e) => (false, format!("Gate check task panicked: {e}")),
        }
    } else {
        (
            false,
            "No worktree path available to run acceptance gates".to_string(),
        )
    };

    slog!(
        "[agents] Server-owned completion for '{story_id}:{agent_name}': gates_passed={gates_passed}"
    );

    let report = CompletionReport {
        summary: "Agent process exited normally".to_string(),
        gates_passed,
        gate_output,
    };

    // Store completion report and set status.
    let tx = {
        let mut lock = match agents.lock() {
            Ok(a) => a,
            Err(_) => return,
        };
        let agent = match lock.get_mut(&key) {
            Some(a) => a,
            None => return,
        };
        agent.completion = Some(report);
        agent.session_id = session_id.clone();
        agent.status = if gates_passed {
            AgentStatus::Completed
        } else {
            AgentStatus::Failed
        };
        agent.completed_at = Some(Instant::now());
        agent.tx.clone()
    };

    // Emit Done so wait_for_agent unblocks.
    let _ = tx.send(AgentEvent::Done {
        story_id: story_id.to_string(),
        agent_name: agent_name.to_string(),
        session_id,
    });

    // Advance the pipeline state machine in a background task.
    // Uses a non-async helper to break the opaque type cycle.
    spawn_pipeline_advance(Arc::clone(agents), port, story_id, agent_name);
}

/// Spawn pipeline advancement as a background task.
///
/// This is a **non-async** function so it does not participate in the opaque
/// type cycle between `start_agent` and `run_server_owned_completion`.
fn spawn_pipeline_advance(
    agents: Arc<Mutex<HashMap<String, StoryAgent>>>,
    port: u16,
    story_id: &str,
    agent_name: &str,
) {
    let sid = story_id.to_string();
    let aname = agent_name.to_string();
    tokio::spawn(async move {
        let pool = AgentPool {
            agents,
            port,
            child_killers: Arc::new(Mutex::new(HashMap::new())),
        };
        pool.run_pipeline_advance_for_completed_agent(&sid, &aname)
            .await;
    });
}

/// Result of a mergemaster merge operation.
#[derive(Debug, Serialize, Clone)]
pub struct MergeReport {
    pub story_id: String,
    pub success: bool,
    pub had_conflicts: bool,
    /// `true` when conflicts were detected but automatically resolved.
    pub conflicts_resolved: bool,
    pub conflict_details: Option<String>,
    pub gates_passed: bool,
    pub gate_output: String,
    pub worktree_cleaned_up: bool,
    pub story_archived: bool,
}

/// Determine the work item type from its ID (new naming: `{N}_{type}_{slug}`).
/// Returns "bug", "spike", or "story".
#[allow(dead_code)]
fn item_type_from_id(item_id: &str) -> &'static str {
    // New format: {digits}_{type}_{slug}
    let after_num = item_id.trim_start_matches(|c: char| c.is_ascii_digit());
    if after_num.starts_with("_bug_") {
        "bug"
    } else if after_num.starts_with("_spike_") {
        "spike"
    } else {
        "story"
    }
}

/// Return the source directory path for a work item (always work/1_upcoming/).
fn item_source_dir(project_root: &Path, _item_id: &str) -> PathBuf {
    project_root.join(".story_kit").join("work").join("1_upcoming")
}

/// Return the done directory path for a work item (always work/5_done/).
fn item_archive_dir(project_root: &Path, _item_id: &str) -> PathBuf {
    project_root.join(".story_kit").join("work").join("5_done")
}

/// Move a work item (story, bug, or spike) from `work/1_upcoming/` to `work/2_current/`.
///
/// Idempotent: if the item is already in `2_current/`, returns Ok without committing.
/// If the item is not found in `1_upcoming/`, logs a warning and returns Ok.
pub fn move_story_to_current(project_root: &Path, story_id: &str) -> Result<(), String> {
    let sk = project_root.join(".story_kit").join("work");
    let current_dir = sk.join("2_current");
    let current_path = current_dir.join(format!("{story_id}.md"));

    if current_path.exists() {
        // Already in 2_current/ — idempotent, nothing to do.
        return Ok(());
    }

    let source_dir = item_source_dir(project_root, story_id);
    let source_path = source_dir.join(format!("{story_id}.md"));

    if !source_path.exists() {
        slog!(
            "[lifecycle] Work item '{story_id}' not found in {}; skipping move to 2_current/",
            source_dir.display()
        );
        return Ok(());
    }

    std::fs::create_dir_all(&current_dir)
        .map_err(|e| format!("Failed to create work/2_current/ directory: {e}"))?;

    std::fs::rename(&source_path, &current_path)
        .map_err(|e| format!("Failed to move '{story_id}' to 2_current/: {e}"))?;

    slog!(
        "[lifecycle] Moved '{story_id}' from {} to work/2_current/",
        source_dir.display()
    );

    Ok(())
}

/// Move a story from `work/2_current/` to `work/5_done/` and auto-commit.
///
/// * If the story is in `2_current/`, it is moved to `5_done/` and committed.
/// * If the story is in `4_merge/`, it is moved to `5_done/` and committed.
/// * If the story is already in `5_done/` or `6_archived/`, this is a no-op (idempotent).
/// * If the story is not found in `2_current/`, `4_merge/`, `5_done/`, or `6_archived/`, an error is returned.
pub fn move_story_to_archived(project_root: &Path, story_id: &str) -> Result<(), String> {
    let sk = project_root.join(".story_kit").join("work");
    let current_path = sk.join("2_current").join(format!("{story_id}.md"));
    let merge_path = sk.join("4_merge").join(format!("{story_id}.md"));
    let done_dir = sk.join("5_done");
    let done_path = done_dir.join(format!("{story_id}.md"));
    let archived_path = sk.join("6_archived").join(format!("{story_id}.md"));

    if done_path.exists() || archived_path.exists() {
        // Already in done or archived — idempotent, nothing to do.
        return Ok(());
    }

    // Check 2_current/ first, then 4_merge/
    let source_path = if current_path.exists() {
        current_path.clone()
    } else if merge_path.exists() {
        merge_path.clone()
    } else {
        return Err(format!(
            "Story '{story_id}' not found in work/2_current/ or work/4_merge/. Cannot accept story."
        ));
    };

    std::fs::create_dir_all(&done_dir)
        .map_err(|e| format!("Failed to create work/5_done/ directory: {e}"))?;
    std::fs::rename(&source_path, &done_path)
        .map_err(|e| format!("Failed to move story '{story_id}' to 5_done/: {e}"))?;

    let from_dir = if source_path == current_path {
        "work/2_current/"
    } else {
        "work/4_merge/"
    };
    slog!("[lifecycle] Moved story '{story_id}' from {from_dir} to work/5_done/");

    Ok(())
}

/// Move a story/bug from `work/2_current/` or `work/3_qa/` to `work/4_merge/`.
///
/// This stages a work item as ready for the mergemaster to pick up and merge into master.
/// Idempotent: if already in `4_merge/`, returns Ok without committing.
pub fn move_story_to_merge(project_root: &Path, story_id: &str) -> Result<(), String> {
    let sk = project_root.join(".story_kit").join("work");
    let current_path = sk.join("2_current").join(format!("{story_id}.md"));
    let qa_path = sk.join("3_qa").join(format!("{story_id}.md"));
    let merge_dir = sk.join("4_merge");
    let merge_path = merge_dir.join(format!("{story_id}.md"));

    if merge_path.exists() {
        // Already in 4_merge/ — idempotent, nothing to do.
        return Ok(());
    }

    // Accept from 2_current/ (manual trigger) or 3_qa/ (pipeline advancement from QA stage).
    let source_path = if current_path.exists() {
        current_path.clone()
    } else if qa_path.exists() {
        qa_path.clone()
    } else {
        return Err(format!(
            "Work item '{story_id}' not found in work/2_current/ or work/3_qa/. Cannot move to 4_merge/."
        ));
    };

    std::fs::create_dir_all(&merge_dir)
        .map_err(|e| format!("Failed to create work/4_merge/ directory: {e}"))?;
    std::fs::rename(&source_path, &merge_path)
        .map_err(|e| format!("Failed to move '{story_id}' to 4_merge/: {e}"))?;

    let from_dir = if source_path == current_path {
        "work/2_current/"
    } else {
        "work/3_qa/"
    };
    slog!("[lifecycle] Moved '{story_id}' from {from_dir} to work/4_merge/");

    Ok(())
}

/// Move a story/bug from `work/2_current/` to `work/3_qa/` and auto-commit.
///
/// This stages a work item for QA review before merging to master.
/// Idempotent: if already in `3_qa/`, returns Ok without committing.
pub fn move_story_to_qa(project_root: &Path, story_id: &str) -> Result<(), String> {
    let sk = project_root.join(".story_kit").join("work");
    let current_path = sk.join("2_current").join(format!("{story_id}.md"));
    let qa_dir = sk.join("3_qa");
    let qa_path = qa_dir.join(format!("{story_id}.md"));

    if qa_path.exists() {
        // Already in 3_qa/ — idempotent, nothing to do.
        return Ok(());
    }

    if !current_path.exists() {
        return Err(format!(
            "Work item '{story_id}' not found in work/2_current/. Cannot move to 3_qa/."
        ));
    }

    std::fs::create_dir_all(&qa_dir)
        .map_err(|e| format!("Failed to create work/3_qa/ directory: {e}"))?;
    std::fs::rename(&current_path, &qa_path)
        .map_err(|e| format!("Failed to move '{story_id}' to 3_qa/: {e}"))?;

    slog!("[lifecycle] Moved '{story_id}' from work/2_current/ to work/3_qa/");

    Ok(())
}

/// Move a bug from `work/2_current/` or `work/1_upcoming/` to `work/5_done/` and auto-commit.
///
/// * If the bug is in `2_current/`, it is moved to `5_done/` and committed.
/// * If the bug is still in `1_upcoming/` (never started), it is moved directly to `5_done/`.
/// * If the bug is already in `5_done/`, this is a no-op (idempotent).
/// * If the bug is not found anywhere, an error is returned.
pub fn close_bug_to_archive(project_root: &Path, bug_id: &str) -> Result<(), String> {
    let sk = project_root.join(".story_kit").join("work");
    let current_path = sk.join("2_current").join(format!("{bug_id}.md"));
    let upcoming_path = sk.join("1_upcoming").join(format!("{bug_id}.md"));
    let archive_dir = item_archive_dir(project_root, bug_id);
    let archive_path = archive_dir.join(format!("{bug_id}.md"));

    if archive_path.exists() {
        return Ok(());
    }

    let source_path = if current_path.exists() {
        current_path.clone()
    } else if upcoming_path.exists() {
        upcoming_path.clone()
    } else {
        return Err(format!(
            "Bug '{bug_id}' not found in work/2_current/ or work/1_upcoming/. Cannot close bug."
        ));
    };

    std::fs::create_dir_all(&archive_dir)
        .map_err(|e| format!("Failed to create work/5_done/ directory: {e}"))?;
    std::fs::rename(&source_path, &archive_path)
        .map_err(|e| format!("Failed to move bug '{bug_id}' to 5_done/: {e}"))?;

    slog!(
        "[lifecycle] Closed bug '{bug_id}' → work/5_done/"
    );

    Ok(())
}

// ── Acceptance-gate helpers ───────────────────────────────────────────────────

/// Detect the base branch for a git worktree by checking common default branch names.
///
/// Tries `master` then `main`; falls back to `"master"` if neither is resolvable.
fn detect_worktree_base_branch(wt_path: &Path) -> String {
    for branch in &["master", "main"] {
        let ok = Command::new("git")
            .args(["rev-parse", "--verify", branch])
            .current_dir(wt_path)
            .output()
            .map(|o| o.status.success())
            .unwrap_or(false);
        if ok {
            return branch.to_string();
        }
    }
    "master".to_string()
}

/// Return `true` if the git worktree at `wt_path` has commits on its current
/// branch that are not present on the base branch (`master` or `main`).
///
/// Used during server startup reconciliation to detect stories whose agent work
/// was committed while the server was offline.
fn worktree_has_committed_work(wt_path: &Path) -> bool {
    let base_branch = detect_worktree_base_branch(wt_path);
    let output = Command::new("git")
        .args(["log", &format!("{base_branch}..HEAD"), "--oneline"])
        .current_dir(wt_path)
        .output();
    match output {
        Ok(out) if out.status.success() => {
            !String::from_utf8_lossy(&out.stdout).trim().is_empty()
        }
        _ => false,
    }
}

/// Check whether the given directory has any uncommitted git changes.
/// Returns `Err` with a descriptive message if there are any.
fn check_uncommitted_changes(path: &Path) -> Result<(), String> {
    let output = Command::new("git")
        .args(["status", "--porcelain"])
        .current_dir(path)
        .output()
        .map_err(|e| format!("Failed to run git status: {e}"))?;

    let stdout = String::from_utf8_lossy(&output.stdout);
    if !stdout.trim().is_empty() {
        return Err(format!(
            "Worktree has uncommitted changes. Please commit all work before \
             the agent exits:\n{stdout}"
        ));
    }
    Ok(())
}

/// Run the project's test suite.
///
/// Uses `script/test` if present, treating it as the canonical single test entry point.
/// Falls back to `cargo nextest run` / `cargo test` when `script/test` is absent.
/// Returns `(tests_passed, output)`.
fn run_project_tests(path: &Path) -> Result<(bool, String), String> {
    let script_test = path.join("script").join("test");
    if script_test.exists() {
        let mut output = String::from("=== script/test ===\n");
        let result = Command::new(&script_test)
            .current_dir(path)
            .output()
            .map_err(|e| format!("Failed to run script/test: {e}"))?;
        let out = format!(
            "{}{}",
            String::from_utf8_lossy(&result.stdout),
            String::from_utf8_lossy(&result.stderr)
        );
        output.push_str(&out);
        output.push('\n');
        return Ok((result.status.success(), output));
    }

    // Fallback: cargo nextest run / cargo test
    let mut output = String::from("=== tests ===\n");
    let (success, test_out) = match Command::new("cargo")
        .args(["nextest", "run"])
        .current_dir(path)
        .output()
    {
        Ok(o) => {
            let combined = format!(
                "{}{}",
                String::from_utf8_lossy(&o.stdout),
                String::from_utf8_lossy(&o.stderr)
            );
            (o.status.success(), combined)
        }
        Err(_) => {
            // nextest not available — fall back to cargo test
            let o = Command::new("cargo")
                .args(["test"])
                .current_dir(path)
                .output()
                .map_err(|e| format!("Failed to run cargo test: {e}"))?;
            let combined = format!(
                "{}{}",
                String::from_utf8_lossy(&o.stdout),
                String::from_utf8_lossy(&o.stderr)
            );
            (o.status.success(), combined)
        }
    };
    output.push_str(&test_out);
    output.push('\n');
    Ok((success, output))
}

/// Run `cargo clippy` and the project test suite (via `script/test` if present,
/// otherwise `cargo nextest run` / `cargo test`) in the given directory.
/// Returns `(gates_passed, combined_output)`.
fn run_acceptance_gates(path: &Path) -> Result<(bool, String), String> {
    let mut all_output = String::new();
    let mut all_passed = true;

    // ── cargo clippy ──────────────────────────────────────────────
    let clippy = Command::new("cargo")
        .args(["clippy", "--all-targets", "--all-features"])
        .current_dir(path)
        .output()
        .map_err(|e| format!("Failed to run cargo clippy: {e}"))?;

    all_output.push_str("=== cargo clippy ===\n");
    let clippy_stdout = String::from_utf8_lossy(&clippy.stdout);
    let clippy_stderr = String::from_utf8_lossy(&clippy.stderr);
    if !clippy_stdout.is_empty() {
        all_output.push_str(&clippy_stdout);
    }
    if !clippy_stderr.is_empty() {
        all_output.push_str(&clippy_stderr);
    }
    all_output.push('\n');

    if !clippy.status.success() {
        all_passed = false;
    }

    // ── tests (script/test if available, else cargo nextest/test) ─
    let (test_success, test_out) = run_project_tests(path)?;
    all_output.push_str(&test_out);
    if !test_success {
        all_passed = false;
    }

    Ok((all_passed, all_output))
}

/// Run `script/test_coverage` in the given directory if the script exists.
///
/// Used as a QA gate before advancing a story from `3_qa/` to `4_merge/`.
/// Returns `(passed, output)`.  If the script does not exist, returns `(true, …)`.
fn run_coverage_gate(path: &Path) -> Result<(bool, String), String> {
    let script = path.join("script").join("test_coverage");
    if !script.exists() {
        return Ok((
            true,
            "script/test_coverage not found; coverage gate skipped.\n".to_string(),
        ));
    }

    let mut output = String::from("=== script/test_coverage ===\n");
    let result = Command::new(&script)
        .current_dir(path)
        .output()
        .map_err(|e| format!("Failed to run script/test_coverage: {e}"))?;

    let combined = format!(
        "{}{}",
        String::from_utf8_lossy(&result.stdout),
        String::from_utf8_lossy(&result.stderr)
    );
    output.push_str(&combined);
    output.push('\n');

    Ok((result.status.success(), output))
}

// ── Mergemaster helpers ───────────────────────────────────────────────────────

/// Result of a squash-merge operation.
struct SquashMergeResult {
    success: bool,
    had_conflicts: bool,
    /// `true` when conflicts were detected but automatically resolved.
    conflicts_resolved: bool,
    conflict_details: Option<String>,
    output: String,
    /// Whether quality gates ran and passed.  `false` when `success` is `false`
    /// due to a gate failure; callers can use this to distinguish gate failures
    /// from merge/commit/FF failures in the `MergeReport`.
    gates_passed: bool,
}

/// Squash-merge a feature branch into the current branch using a temporary
/// merge-queue worktree. This avoids the race condition where the filesystem
/// watcher auto-commits conflict markers to master.
///
/// **Flow:**
/// 1. Create a temporary `merge-queue/{story_id}` branch at current HEAD.
/// 2. Create a temporary worktree for that branch.
/// 3. Run `git merge --squash` in the temporary worktree (not the main worktree).
/// 4. If conflicts arise, attempt automatic resolution for simple additive cases.
/// 5. If clean (or resolved), commit in the temp worktree.
/// 6. Run quality gates **in the merge worktree** before touching master.
/// 7. If gates pass: fast-forward master to the merge-queue commit.
/// 8. Clean up the temporary worktree and branch.
fn run_squash_merge(
    project_root: &Path,
    branch: &str,
    story_id: &str,
) -> Result<SquashMergeResult, String> {
    let mut all_output = String::new();
    let merge_branch = format!("merge-queue/{story_id}");
    let merge_wt_path = project_root
        .join(".story_kit")
        .join("merge_workspace");

    // Ensure we start clean: remove any leftover merge workspace.
    cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);

    // ── Create merge-queue branch at current HEAD ─────────────────
    all_output.push_str(&format!(
        "=== Creating merge-queue branch '{merge_branch}' ===\n"
    ));
    let create_branch = Command::new("git")
        .args(["branch", &merge_branch])
        .current_dir(project_root)
        .output()
        .map_err(|e| format!("Failed to create merge-queue branch: {e}"))?;
    if !create_branch.status.success() {
        let stderr = String::from_utf8_lossy(&create_branch.stderr);
        all_output.push_str(&format!("Branch creation failed: {stderr}\n"));
        return Err(format!("Failed to create merge-queue branch: {stderr}"));
    }

    // ── Create temporary worktree ─────────────────────────────────
    all_output.push_str("=== Creating temporary merge worktree ===\n");
    let wt_str = merge_wt_path.to_string_lossy().to_string();
    let create_wt = Command::new("git")
        .args(["worktree", "add", &wt_str, &merge_branch])
        .current_dir(project_root)
        .output()
        .map_err(|e| format!("Failed to create merge worktree: {e}"))?;
    if !create_wt.status.success() {
        let stderr = String::from_utf8_lossy(&create_wt.stderr);
        all_output.push_str(&format!("Worktree creation failed: {stderr}\n"));
        cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);
        return Err(format!("Failed to create merge worktree: {stderr}"));
    }

    // ── Squash-merge in the temporary worktree ────────────────────
    all_output.push_str(&format!("=== git merge --squash {branch} ===\n"));
    let merge = Command::new("git")
        .args(["merge", "--squash", branch])
        .current_dir(&merge_wt_path)
        .output()
        .map_err(|e| format!("Failed to run git merge: {e}"))?;

    let merge_stdout = String::from_utf8_lossy(&merge.stdout).to_string();
    let merge_stderr = String::from_utf8_lossy(&merge.stderr).to_string();
    all_output.push_str(&merge_stdout);
    all_output.push_str(&merge_stderr);
    all_output.push('\n');

    let mut had_conflicts = false;
    let mut conflicts_resolved = false;
    let mut conflict_details: Option<String> = None;

    if !merge.status.success() {
        had_conflicts = true;
        all_output.push_str("=== Conflicts detected, attempting auto-resolution ===\n");

        // Try to automatically resolve simple conflicts.
        match try_resolve_conflicts(&merge_wt_path) {
            Ok((resolved, resolution_log)) => {
                all_output.push_str(&resolution_log);
                if resolved {
                    conflicts_resolved = true;
                    all_output
                        .push_str("=== All conflicts resolved automatically ===\n");
                } else {
                    // Could not resolve — abort, clean up, and report.
                    let details = format!(
                        "Merge conflicts in branch '{branch}':\n{merge_stdout}{merge_stderr}\n{resolution_log}"
                    );
                    conflict_details = Some(details);
                    all_output
                        .push_str("=== Unresolvable conflicts, aborting merge ===\n");
                    cleanup_merge_workspace(
                        project_root,
                        &merge_wt_path,
                        &merge_branch,
                    );
                    return Ok(SquashMergeResult {
                        success: false,
                        had_conflicts: true,
                        conflicts_resolved: false,
                        conflict_details,
                        output: all_output,
                        gates_passed: false,
                    });
                }
            }
            Err(e) => {
                all_output.push_str(&format!("Auto-resolution error: {e}\n"));
                cleanup_merge_workspace(
                    project_root,
                    &merge_wt_path,
                    &merge_branch,
                );
                return Ok(SquashMergeResult {
                    success: false,
                    had_conflicts: true,
                    conflicts_resolved: false,
                    conflict_details: Some(format!(
                        "Merge conflicts in branch '{branch}' (auto-resolution failed: {e}):\n{merge_stdout}{merge_stderr}"
                    )),
                    output: all_output,
                    gates_passed: false,
                });
            }
        }
    }

    // ── Commit in the temporary worktree ──────────────────────────
    all_output.push_str("=== git commit ===\n");
    let commit_msg = format!("story-kit: merge {story_id}");
    let commit = Command::new("git")
        .args(["commit", "-m", &commit_msg])
        .current_dir(&merge_wt_path)
        .output()
        .map_err(|e| format!("Failed to run git commit: {e}"))?;

    let commit_stdout = String::from_utf8_lossy(&commit.stdout).to_string();
    let commit_stderr = String::from_utf8_lossy(&commit.stderr).to_string();
    all_output.push_str(&commit_stdout);
    all_output.push_str(&commit_stderr);
    all_output.push('\n');

    if !commit.status.success() {
        // Nothing to commit (e.g. empty diff) — treat as success.
        if commit_stderr.contains("nothing to commit")
            || commit_stdout.contains("nothing to commit")
        {
            cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);
            return Ok(SquashMergeResult {
                success: true,
                had_conflicts,
                conflicts_resolved,
                conflict_details,
                output: all_output,
                gates_passed: true,
            });
        }
        cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);
        return Ok(SquashMergeResult {
            success: false,
            had_conflicts,
            conflicts_resolved,
            conflict_details,
            output: all_output,
            gates_passed: false,
        });
    }

    // ── Install frontend dependencies for quality gates ──────────────
    let frontend_dir_for_install = merge_wt_path.join("frontend");
    if frontend_dir_for_install.exists() {
        // Ensure frontend/dist/ exists so cargo clippy (RustEmbed) can compile
        // even before `pnpm build` has run.
        let dist_dir = frontend_dir_for_install.join("dist");
        std::fs::create_dir_all(&dist_dir)
            .map_err(|e| format!("Failed to create frontend/dist: {e}"))?;

        all_output.push_str("=== pnpm install (merge worktree) ===\n");
        let pnpm_install = Command::new("pnpm")
            .args(["install"])
            .current_dir(&frontend_dir_for_install)
            .output()
            .map_err(|e| format!("Failed to run pnpm install: {e}"))?;

        let install_out = format!(
            "{}{}",
            String::from_utf8_lossy(&pnpm_install.stdout),
            String::from_utf8_lossy(&pnpm_install.stderr)
        );
        all_output.push_str(&install_out);
        all_output.push('\n');

        if !pnpm_install.status.success() {
            all_output.push_str("=== pnpm install FAILED — aborting merge ===\n");
            cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);
            return Ok(SquashMergeResult {
                success: false,
                had_conflicts,
                conflicts_resolved,
                conflict_details,
                output: all_output,
                gates_passed: false,
            });
        }
    }

    // ── Install frontend dependencies for quality gates ──────────
    let frontend_dir = merge_wt_path.join("frontend");
    if frontend_dir.exists() {
        // Ensure frontend/dist exists so RustEmbed (cargo clippy) doesn't fail
        // even before pnpm build runs.
        let dist_dir = frontend_dir.join("dist");
        if !dist_dir.exists() {
            let _ = std::fs::create_dir_all(&dist_dir);
        }

        all_output.push_str("=== pnpm install (merge worktree) ===\n");
        let pnpm_install = Command::new("pnpm")
            .args(["install", "--frozen-lockfile"])
            .current_dir(&frontend_dir)
            .output()
            .map_err(|e| format!("Failed to run pnpm install: {e}"))?;

        let install_out = format!(
            "{}{}",
            String::from_utf8_lossy(&pnpm_install.stdout),
            String::from_utf8_lossy(&pnpm_install.stderr)
        );
        all_output.push_str(&install_out);
        all_output.push('\n');

        if !pnpm_install.status.success() {
            all_output.push_str(
                "=== pnpm install FAILED — aborting merge, master unchanged ===\n",
            );
            cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);
            return Ok(SquashMergeResult {
                success: false,
                had_conflicts,
                conflicts_resolved,
                conflict_details,
                output: all_output,
                gates_passed: false,
            });
        }
    }

    // ── Quality gates in merge workspace (BEFORE fast-forward) ────
    // Run gates in the merge worktree so that failures abort before master moves.
    all_output.push_str("=== Running quality gates before fast-forward ===\n");
    match run_merge_quality_gates(&merge_wt_path) {
        Ok((true, gate_out)) => {
            all_output.push_str(&gate_out);
            all_output.push('\n');
            all_output.push_str("=== Quality gates passed ===\n");
        }
        Ok((false, gate_out)) => {
            all_output.push_str(&gate_out);
            all_output.push('\n');
            all_output
                .push_str("=== Quality gates FAILED — aborting fast-forward, master unchanged ===\n");
            cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);
            return Ok(SquashMergeResult {
                success: false,
                had_conflicts,
                conflicts_resolved,
                conflict_details,
                output: all_output,
                gates_passed: false,
            });
        }
        Err(e) => {
            all_output.push_str(&format!("Gate check error: {e}\n"));
            cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);
            return Ok(SquashMergeResult {
                success: false,
                had_conflicts,
                conflicts_resolved,
                conflict_details,
                output: all_output,
                gates_passed: false,
            });
        }
    }

    // ── Fast-forward master to the merge-queue commit ─────────────
    all_output.push_str(&format!(
        "=== Fast-forwarding master to {merge_branch} ===\n"
    ));
    let ff = Command::new("git")
        .args(["merge", "--ff-only", &merge_branch])
        .current_dir(project_root)
        .output()
        .map_err(|e| format!("Failed to fast-forward master: {e}"))?;

    let ff_stdout = String::from_utf8_lossy(&ff.stdout).to_string();
    let ff_stderr = String::from_utf8_lossy(&ff.stderr).to_string();
    all_output.push_str(&ff_stdout);
    all_output.push_str(&ff_stderr);
    all_output.push('\n');

    if !ff.status.success() {
        all_output.push_str("=== Fast-forward failed — master may have diverged ===\n");
        cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);
        return Ok(SquashMergeResult {
            success: false,
            had_conflicts,
            conflicts_resolved,
            conflict_details: Some(format!(
                "Fast-forward to merge-queue failed (master diverged?):\n{ff_stderr}"
            )),
            output: all_output,
            gates_passed: true,
        });
    }

    // ── Clean up ──────────────────────────────────────────────────
    cleanup_merge_workspace(project_root, &merge_wt_path, &merge_branch);
    all_output.push_str("=== Merge-queue cleanup complete ===\n");

    Ok(SquashMergeResult {
        success: true,
        had_conflicts,
        conflicts_resolved,
        conflict_details,
        output: all_output,
        gates_passed: true,
    })
}

/// Remove the temporary merge worktree and branch. Best-effort — errors are
/// silently ignored because this is cleanup code.
fn cleanup_merge_workspace(
    project_root: &Path,
    merge_wt_path: &Path,
    merge_branch: &str,
) {
    let wt_str = merge_wt_path.to_string_lossy().to_string();
    let _ = Command::new("git")
        .args(["worktree", "remove", "--force", &wt_str])
        .current_dir(project_root)
        .output();
    let _ = Command::new("git")
        .args(["branch", "-D", merge_branch])
        .current_dir(project_root)
        .output();
}

/// Attempt to automatically resolve merge conflicts in the given worktree.
///
/// Finds all conflicted files and tries [`resolve_simple_conflicts`] on each.
/// If **all** conflicts can be resolved, stages the resolved files and returns
/// `Ok((true, log))`. If any file has a complex conflict that cannot be
/// auto-resolved, returns `Ok((false, log))` without staging anything.
fn try_resolve_conflicts(worktree: &Path) -> Result<(bool, String), String> {
    let mut log = String::new();

    // List conflicted files.
    let ls = Command::new("git")
        .args(["diff", "--name-only", "--diff-filter=U"])
        .current_dir(worktree)
        .output()
        .map_err(|e| format!("Failed to list conflicted files: {e}"))?;

    let file_list = String::from_utf8_lossy(&ls.stdout);
    let conflicted_files: Vec<&str> =
        file_list.lines().filter(|l| !l.is_empty()).collect();

    if conflicted_files.is_empty() {
        log.push_str("No conflicted files found (conflict may be index-only).\n");
        return Ok((false, log));
    }

    log.push_str(&format!(
        "Conflicted files ({}):\n",
        conflicted_files.len()
    ));
    for f in &conflicted_files {
        log.push_str(&format!("  - {f}\n"));
    }

    // First pass: check that all files can be resolved before touching any.
    let mut resolutions: Vec<(&str, String)> = Vec::new();
    for file in &conflicted_files {
        let file_path = worktree.join(file);
        let content = std::fs::read_to_string(&file_path)
            .map_err(|e| format!("Failed to read conflicted file '{file}': {e}"))?;

        match resolve_simple_conflicts(&content) {
            Some(resolved) => {
                log.push_str(&format!("  [auto-resolve] {file}\n"));
                resolutions.push((file, resolved));
            }
            None => {
                log.push_str(&format!(
                    "  [COMPLEX — cannot auto-resolve] {file}\n"
                ));
                return Ok((false, log));
            }
        }
    }

    // Second pass: write resolved content and stage.
    for (file, resolved) in &resolutions {
        let file_path = worktree.join(file);
        std::fs::write(&file_path, resolved)
            .map_err(|e| format!("Failed to write resolved file '{file}': {e}"))?;

        let add = Command::new("git")
            .args(["add", file])
            .current_dir(worktree)
            .output()
            .map_err(|e| format!("Failed to stage resolved file '{file}': {e}"))?;
        if !add.status.success() {
            return Err(format!(
                "git add failed for '{file}': {}",
                String::from_utf8_lossy(&add.stderr)
            ));
        }
    }

    Ok((true, log))
}

/// Try to resolve simple additive merge conflicts in a file's content.
///
/// A conflict is considered "simple additive" when both sides add new content
/// at the same location without modifying existing lines. In that case we keep
/// both additions (ours first, then theirs).
///
/// Returns `Some(resolved)` if all conflict blocks in the file are simple, or
/// `None` if any block is too complex to auto-resolve.
fn resolve_simple_conflicts(content: &str) -> Option<String> {
    // Quick check: if there are no conflict markers at all, nothing to do.
    if !content.contains("<<<<<<<") {
        return Some(content.to_string());
    }

    let mut result = String::new();
    let mut lines = content.lines().peekable();

    while let Some(line) = lines.next() {
        if line.starts_with("<<<<<<<") {
            // Collect the "ours" side (between <<<<<<< and =======).
            let mut ours = Vec::new();
            let mut found_separator = false;
            for next_line in lines.by_ref() {
                if next_line.starts_with("=======") {
                    found_separator = true;
                    break;
                }
                ours.push(next_line);
            }
            if !found_separator {
                return None; // Malformed conflict block.
            }

            // Collect the "theirs" side (between ======= and >>>>>>>).
            let mut theirs = Vec::new();
            let mut found_end = false;
            for next_line in lines.by_ref() {
                if next_line.starts_with(">>>>>>>") {
                    found_end = true;
                    break;
                }
                theirs.push(next_line);
            }
            if !found_end {
                return None; // Malformed conflict block.
            }

            // Both sides must be non-empty additions to be considered simple.
            // If either side is empty, it means one side deleted something — complex.
            if ours.is_empty() && theirs.is_empty() {
                // Both empty — nothing to add, skip.
                continue;
            }

            // Accept both: ours first, then theirs.
            for l in &ours {
                result.push_str(l);
                result.push('\n');
            }
            for l in &theirs {
                result.push_str(l);
                result.push('\n');
            }
        } else {
            result.push_str(line);
            result.push('\n');
        }
    }

    // Preserve trailing newline consistency: if original ended without
    // newline, strip the trailing one we added.
    if !content.ends_with('\n') && result.ends_with('\n') {
        result.pop();
    }

    Some(result)
}

/// Run quality gates in the project root after a successful merge.
///
/// Runs: cargo clippy and tests if Cargo.toml is present, script/test if present, and pnpm gates
/// if frontend/ exists.  Sections are skipped when the corresponding project artefacts are absent.
/// Returns `(gates_passed, combined_output)`.
fn run_merge_quality_gates(project_root: &Path) -> Result<(bool, String), String> {
    let mut all_output = String::new();
    let mut all_passed = true;

    let cargo_toml = project_root.join("Cargo.toml");
    let script_test = project_root.join("script").join("test");

    // ── cargo clippy (only when a Rust project is present) ────────
    if cargo_toml.exists() {
        let clippy = Command::new("cargo")
            .args(["clippy", "--all-targets", "--all-features"])
            .current_dir(project_root)
            .output()
            .map_err(|e| format!("Failed to run cargo clippy: {e}"))?;

        all_output.push_str("=== cargo clippy ===\n");
        let clippy_out = format!(
            "{}{}",
            String::from_utf8_lossy(&clippy.stdout),
            String::from_utf8_lossy(&clippy.stderr)
        );
        all_output.push_str(&clippy_out);
        all_output.push('\n');

        if !clippy.status.success() {
            all_passed = false;
        }
    }

    // ── tests (script/test if available, else cargo nextest/test) ─
    // Only run when there is something to run: a script/test entrypoint or a
    // Rust project (Cargo.toml).  Skipping avoids spurious failures in
    // environments that have neither (e.g. test temp-dirs).
    if script_test.exists() || cargo_toml.exists() {
        let (test_success, test_out) = run_project_tests(project_root)?;
        all_output.push_str(&test_out);
        if !test_success {
            all_passed = false;
        }
    }

    // ── pnpm build (if frontend/ directory exists) ────────────────
    // pnpm test is handled by script/test when present; only run it here as
    // a standalone fallback when there is no script/test.
    let frontend_dir = project_root.join("frontend");
    if frontend_dir.exists() {
        all_output.push_str("=== pnpm build ===\n");
        let pnpm_build = Command::new("pnpm")
            .args(["run", "build"])
            .current_dir(&frontend_dir)
            .output()
            .map_err(|e| format!("Failed to run pnpm build: {e}"))?;

        let build_out = format!(
            "{}{}",
            String::from_utf8_lossy(&pnpm_build.stdout),
            String::from_utf8_lossy(&pnpm_build.stderr)
        );
        all_output.push_str(&build_out);
        all_output.push('\n');

        if !pnpm_build.status.success() {
            all_passed = false;
        }

        // Only run pnpm test separately when script/test is absent (it would
        // already cover frontend tests in that case).
        let script_test = project_root.join("script").join("test");
        if !script_test.exists() {
            all_output.push_str("=== pnpm test ===\n");
            let pnpm_test = Command::new("pnpm")
                .args(["test", "--run"])
                .current_dir(&frontend_dir)
                .output()
                .map_err(|e| format!("Failed to run pnpm test: {e}"))?;

            let pnpm_test_out = format!(
                "{}{}",
                String::from_utf8_lossy(&pnpm_test.stdout),
                String::from_utf8_lossy(&pnpm_test.stderr)
            );
            all_output.push_str(&pnpm_test_out);
            all_output.push('\n');

            if !pnpm_test.status.success() {
                all_passed = false;
            }
        }
    }

    Ok((all_passed, all_output))
}

/// Spawn claude agent in a PTY and stream events through the broadcast channel.
#[allow(clippy::too_many_arguments)]
async fn run_agent_pty_streaming(
    story_id: &str,
    agent_name: &str,
    command: &str,
    args: &[String],
    prompt: &str,
    cwd: &str,
    tx: &broadcast::Sender<AgentEvent>,
    event_log: &Arc<Mutex<Vec<AgentEvent>>>,
    log_writer: Option<Arc<Mutex<AgentLogWriter>>>,
    inactivity_timeout_secs: u64,
    child_killers: Arc<Mutex<HashMap<String, Box<dyn ChildKiller + Send + Sync>>>>,
) -> Result<Option<String>, String> {
    let sid = story_id.to_string();
    let aname = agent_name.to_string();
    let cmd = command.to_string();
    let args = args.to_vec();
    let prompt = prompt.to_string();
    let cwd = cwd.to_string();
    let tx = tx.clone();
    let event_log = event_log.clone();

    tokio::task::spawn_blocking(move || {
        run_agent_pty_blocking(
            &sid,
            &aname,
            &cmd,
            &args,
            &prompt,
            &cwd,
            &tx,
            &event_log,
            log_writer.as_deref(),
            inactivity_timeout_secs,
            &child_killers,
        )
    })
    .await
    .map_err(|e| format!("Agent task panicked: {e}"))?
}

/// Dispatch a `stream_event` from Claude Code's `--include-partial-messages` output.
///
/// Extracts `thinking_delta` and `text_delta` from `content_block_delta` events
/// and routes them as `AgentEvent::Thinking` and `AgentEvent::Output` respectively.
/// This ensures thinking traces flow through the dedicated `ThinkingBlock` UI
/// component rather than appearing as unbounded regular output.
fn handle_agent_stream_event(
    event: &serde_json::Value,
    story_id: &str,
    agent_name: &str,
    tx: &broadcast::Sender<AgentEvent>,
    event_log: &Mutex<Vec<AgentEvent>>,
    log_writer: Option<&Mutex<AgentLogWriter>>,
) {
    let event_type = event.get("type").and_then(|t| t.as_str()).unwrap_or("");

    if event_type == "content_block_delta"
        && let Some(delta) = event.get("delta")
    {
        let delta_type = delta.get("type").and_then(|t| t.as_str()).unwrap_or("");
        match delta_type {
            "thinking_delta" => {
                if let Some(thinking) = delta.get("thinking").and_then(|t| t.as_str()) {
                    emit_event(
                        AgentEvent::Thinking {
                            story_id: story_id.to_string(),
                            agent_name: agent_name.to_string(),
                            text: thinking.to_string(),
                        },
                        tx,
                        event_log,
                        log_writer,
                    );
                }
            }
            "text_delta" => {
                if let Some(text) = delta.get("text").and_then(|t| t.as_str()) {
                    emit_event(
                        AgentEvent::Output {
                            story_id: story_id.to_string(),
                            agent_name: agent_name.to_string(),
                            text: text.to_string(),
                        },
                        tx,
                        event_log,
                        log_writer,
                    );
                }
            }
            _ => {}
        }
    }
}

/// Helper to send an event to broadcast, event log, and optional persistent log file.
fn emit_event(
    event: AgentEvent,
    tx: &broadcast::Sender<AgentEvent>,
    event_log: &Mutex<Vec<AgentEvent>>,
    log_writer: Option<&Mutex<AgentLogWriter>>,
) {
    if let Ok(mut log) = event_log.lock() {
        log.push(event.clone());
    }
    if let Some(writer) = log_writer
        && let Ok(mut w) = writer.lock()
        && let Err(e) = w.write_event(&event)
    {
        eprintln!("[agent_log] Failed to write event to log file: {e}");
    }
    let _ = tx.send(event);
}

#[allow(clippy::too_many_arguments)]
fn run_agent_pty_blocking(
    story_id: &str,
    agent_name: &str,
    command: &str,
    args: &[String],
    prompt: &str,
    cwd: &str,
    tx: &broadcast::Sender<AgentEvent>,
    event_log: &Mutex<Vec<AgentEvent>>,
    log_writer: Option<&Mutex<AgentLogWriter>>,
    inactivity_timeout_secs: u64,
    child_killers: &Arc<Mutex<HashMap<String, Box<dyn ChildKiller + Send + Sync>>>>,
) -> Result<Option<String>, String> {
    let pty_system = native_pty_system();

    let pair = pty_system
        .openpty(PtySize {
            rows: 50,
            cols: 200,
            pixel_width: 0,
            pixel_height: 0,
        })
        .map_err(|e| format!("Failed to open PTY: {e}"))?;

    let mut cmd = CommandBuilder::new(command);

    // -p <prompt> must come first
    cmd.arg("-p");
    cmd.arg(prompt);

    // Add configured args (e.g., --directory /path/to/worktree, --model, etc.)
    for arg in args {
        cmd.arg(arg);
    }

    cmd.arg("--output-format");
    cmd.arg("stream-json");
    cmd.arg("--verbose");
    // Enable partial streaming so we receive thinking_delta and text_delta
    // events in real-time, rather than only complete assistant events.
    // Without this, thinking traces may not appear in the structured output
    // and instead leak as unstructured PTY text.
    cmd.arg("--include-partial-messages");

    // Supervised agents don't need interactive permission prompts
    cmd.arg("--permission-mode");
    cmd.arg("bypassPermissions");

    cmd.cwd(cwd);
    cmd.env("NO_COLOR", "1");

    // Allow spawning Claude Code from within a Claude Code session
    cmd.env_remove("CLAUDECODE");
    cmd.env_remove("CLAUDE_CODE_ENTRYPOINT");

    slog!("[agent:{story_id}:{agent_name}] Spawning {command} in {cwd} with args: {args:?}");

    let mut child = pair
        .slave
        .spawn_command(cmd)
        .map_err(|e| format!("Failed to spawn agent for {story_id}:{agent_name}: {e}"))?;

    // Register the child killer so that kill_all_children() / stop_agent() can
    // terminate this process on server shutdown, even if the blocking thread
    // cannot be interrupted.  The ChildKillerGuard deregisters on function exit.
    let killer_key = composite_key(story_id, agent_name);
    {
        let killer = child.clone_killer();
        if let Ok(mut killers) = child_killers.lock() {
            killers.insert(killer_key.clone(), killer);
        }
    }
    let _killer_guard = ChildKillerGuard {
        killers: Arc::clone(child_killers),
        key: killer_key,
    };

    drop(pair.slave);

    let reader = pair
        .master
        .try_clone_reader()
        .map_err(|e| format!("Failed to clone PTY reader: {e}"))?;

    drop(pair.master);

    // Spawn a reader thread to collect PTY output lines.
    // We use a channel so the main thread can apply an inactivity deadline
    // via recv_timeout: if no output arrives within the configured window
    // the process is killed and the agent is marked Failed.
    let (line_tx, line_rx) = std::sync::mpsc::channel::<std::io::Result<String>>();
    std::thread::spawn(move || {
        let buf_reader = BufReader::new(reader);
        for line in buf_reader.lines() {
            if line_tx.send(line).is_err() {
                break;
            }
        }
    });

    let timeout_dur = if inactivity_timeout_secs > 0 {
        Some(std::time::Duration::from_secs(inactivity_timeout_secs))
    } else {
        None
    };

    let mut session_id: Option<String> = None;

    loop {
        let recv_result = match timeout_dur {
            Some(dur) => line_rx.recv_timeout(dur),
            None => line_rx
                .recv()
                .map_err(|_| std::sync::mpsc::RecvTimeoutError::Disconnected),
        };

        let line = match recv_result {
            Ok(Ok(l)) => l,
            Ok(Err(_)) => {
                // IO error reading from PTY — treat as EOF.
                break;
            }
            Err(std::sync::mpsc::RecvTimeoutError::Disconnected) => {
                // Reader thread exited (EOF from PTY).
                break;
            }
            Err(std::sync::mpsc::RecvTimeoutError::Timeout) => {
                slog_warn!(
                    "[agent:{story_id}:{agent_name}] Inactivity timeout after \
                     {inactivity_timeout_secs}s with no output. Killing process."
                );
                let _ = child.kill();
                let _ = child.wait();
                return Err(format!(
                    "Agent inactivity timeout: no output received for {inactivity_timeout_secs}s"
                ));
            }
        };

        let trimmed = line.trim();
        if trimmed.is_empty() {
            continue;
        }

        // Try to parse as JSON
        let json: serde_json::Value = match serde_json::from_str(trimmed) {
            Ok(j) => j,
            Err(_) => {
                // Non-JSON output (terminal escapes etc.) — send as raw output
                emit_event(
                    AgentEvent::Output {
                        story_id: story_id.to_string(),
                        agent_name: agent_name.to_string(),
                        text: trimmed.to_string(),
                    },
                    tx,
                    event_log,
                    log_writer,
                );
                continue;
            }
        };

        let event_type = json.get("type").and_then(|t| t.as_str()).unwrap_or("");

        match event_type {
            "system" => {
                session_id = json
                    .get("session_id")
                    .and_then(|s| s.as_str())
                    .map(|s| s.to_string());
            }
            // With --include-partial-messages, thinking and text arrive
            // incrementally via stream_event → content_block_delta. Handle
            // them here for real-time streaming to the frontend.
            "stream_event" => {
                if let Some(event) = json.get("event") {
                    handle_agent_stream_event(
                        event,
                        story_id,
                        agent_name,
                        tx,
                        event_log,
                        log_writer,
                    );
                }
            }
            // Complete assistant events are skipped for content extraction
            // because thinking and text already arrived via stream_event.
            // The raw JSON is still forwarded as AgentJson below.
            "assistant" | "user" | "result" => {}
            _ => {}
        }

        // Forward all JSON events
        emit_event(
            AgentEvent::AgentJson {
                story_id: story_id.to_string(),
                agent_name: agent_name.to_string(),
                data: json,
            },
            tx,
            event_log,
            log_writer,
        );
    }

    let _ = child.kill();
    let _ = child.wait();

    slog!(
        "[agent:{story_id}:{agent_name}] Done. Session: {:?}",
        session_id
    );

    Ok(session_id)
}

#[cfg(test)]
mod tests {
    use super::*;

    #[tokio::test]
    async fn wait_for_agent_returns_immediately_if_completed() {
        let pool = AgentPool::new(3001);
        pool.inject_test_agent("s1", "bot", AgentStatus::Completed);

        let info = pool.wait_for_agent("s1", "bot", 1000).await.unwrap();
        assert_eq!(info.status, AgentStatus::Completed);
        assert_eq!(info.story_id, "s1");
        assert_eq!(info.agent_name, "bot");
    }

    #[tokio::test]
    async fn wait_for_agent_returns_immediately_if_failed() {
        let pool = AgentPool::new(3001);
        pool.inject_test_agent("s2", "bot", AgentStatus::Failed);

        let info = pool.wait_for_agent("s2", "bot", 1000).await.unwrap();
        assert_eq!(info.status, AgentStatus::Failed);
    }

    #[tokio::test]
    async fn wait_for_agent_completes_on_done_event() {
        let pool = AgentPool::new(3001);
        let tx = pool.inject_test_agent("s3", "bot", AgentStatus::Running);

        // Send Done event after a short delay
        let tx_clone = tx.clone();
        tokio::spawn(async move {
            tokio::time::sleep(std::time::Duration::from_millis(50)).await;
            // Mark status via event; real code also updates the map, but for
            // this unit test the map entry stays Running — we verify the
            // wait loop reacts to the event.
            let _ = tx_clone.send(AgentEvent::Done {
                story_id: "s3".to_string(),
                agent_name: "bot".to_string(),
                session_id: Some("sess-abc".to_string()),
            });
        });

        let info = pool.wait_for_agent("s3", "bot", 2000).await.unwrap();
        // Status comes from the map entry (Running in this unit test)
        // — the important thing is that wait_for_agent returned without timing out.
        assert_eq!(info.story_id, "s3");
    }

    #[tokio::test]
    async fn wait_for_agent_times_out() {
        let pool = AgentPool::new(3001);
        pool.inject_test_agent("s4", "bot", AgentStatus::Running);

        let result = pool.wait_for_agent("s4", "bot", 50).await;
        assert!(result.is_err());
        let msg = result.unwrap_err();
        assert!(msg.contains("Timed out"), "unexpected message: {msg}");
    }

    #[tokio::test]
    async fn wait_for_agent_errors_for_nonexistent() {
        let pool = AgentPool::new(3001);
        let result = pool.wait_for_agent("no_story", "no_bot", 100).await;
        assert!(result.is_err());
    }

    #[tokio::test]
    async fn wait_for_agent_completes_on_stopped_status_event() {
        let pool = AgentPool::new(3001);
        let tx = pool.inject_test_agent("s5", "bot", AgentStatus::Running);

        let tx_clone = tx.clone();
        tokio::spawn(async move {
            tokio::time::sleep(std::time::Duration::from_millis(30)).await;
            let _ = tx_clone.send(AgentEvent::Status {
                story_id: "s5".to_string(),
                agent_name: "bot".to_string(),
                status: "stopped".to_string(),
            });
        });

        let info = pool.wait_for_agent("s5", "bot", 2000).await.unwrap();
        assert_eq!(info.story_id, "s5");
    }

    // ── report_completion tests ────────────────────────────────────

    #[tokio::test]
    async fn report_completion_rejects_nonexistent_agent() {
        let pool = AgentPool::new(3001);
        let result = pool
            .report_completion("no_story", "no_bot", "done")
            .await;
        assert!(result.is_err());
        let msg = result.unwrap_err();
        assert!(msg.contains("No agent"), "unexpected: {msg}");
    }

    #[tokio::test]
    async fn report_completion_rejects_non_running_agent() {
        let pool = AgentPool::new(3001);
        pool.inject_test_agent("s6", "bot", AgentStatus::Completed);

        let result = pool.report_completion("s6", "bot", "done").await;
        assert!(result.is_err());
        let msg = result.unwrap_err();
        assert!(
            msg.contains("not running"),
            "expected 'not running' in: {msg}"
        );
    }

    #[tokio::test]
    async fn report_completion_rejects_dirty_worktree() {
        use std::fs;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();

        // Init a real git repo and make an initial commit
        Command::new("git")
            .args(["init"])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "--allow-empty", "-m", "init"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Write an uncommitted file
        fs::write(repo.join("dirty.txt"), "not committed").unwrap();

        let pool = AgentPool::new(3001);
        pool.inject_test_agent_with_path("s7", "bot", AgentStatus::Running, repo.to_path_buf());

        let result = pool.report_completion("s7", "bot", "done").await;
        assert!(result.is_err());
        let msg = result.unwrap_err();
        assert!(
            msg.contains("uncommitted"),
            "expected 'uncommitted' in: {msg}"
        );
    }

    // ── server-owned completion tests ───────────────────────────────────────────

    #[tokio::test]
    async fn server_owned_completion_skips_when_already_completed() {
        let pool = AgentPool::new(3001);
        let report = CompletionReport {
            summary: "Already done".to_string(),
            gates_passed: true,
            gate_output: String::new(),
        };
        pool.inject_test_agent_with_completion(
            "s10",
            "coder-1",
            AgentStatus::Completed,
            PathBuf::from("/tmp/nonexistent"),
            report,
        );

        // Subscribe before calling so we can check if Done event was emitted.
        let mut rx = pool.subscribe("s10", "coder-1").unwrap();

        run_server_owned_completion(&pool.agents, pool.port, "s10", "coder-1", Some("sess-1".to_string()))
            .await;

        // Status should remain Completed (unchanged) — no gate re-run.
        let agents = pool.agents.lock().unwrap();
        let key = composite_key("s10", "coder-1");
        let agent = agents.get(&key).unwrap();
        assert_eq!(agent.status, AgentStatus::Completed);
        // Summary should still be the original, not overwritten.
        assert_eq!(
            agent.completion.as_ref().unwrap().summary,
            "Already done"
        );
        drop(agents);

        // No Done event should have been emitted.
        assert!(
            rx.try_recv().is_err(),
            "should not emit Done when completion already exists"
        );
    }

    #[tokio::test]
    async fn server_owned_completion_runs_gates_on_clean_worktree() {
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        let pool = AgentPool::new(3001);
        pool.inject_test_agent_with_path(
            "s11",
            "coder-1",
            AgentStatus::Running,
            repo.to_path_buf(),
        );

        let mut rx = pool.subscribe("s11", "coder-1").unwrap();

        run_server_owned_completion(&pool.agents, pool.port, "s11", "coder-1", Some("sess-2".to_string()))
            .await;

        // Completion report should exist (gates were run, though they may fail
        // because this is not a real Cargo project).
        let agents = pool.agents.lock().unwrap();
        let key = composite_key("s11", "coder-1");
        let agent = agents.get(&key).unwrap();
        assert!(
            agent.completion.is_some(),
            "completion report should be created"
        );
        assert_eq!(
            agent.completion.as_ref().unwrap().summary,
            "Agent process exited normally"
        );
        // Session ID should be stored.
        assert_eq!(agent.session_id, Some("sess-2".to_string()));
        // Status should be terminal (Completed or Failed depending on gate results).
        assert!(
            agent.status == AgentStatus::Completed || agent.status == AgentStatus::Failed,
            "status should be terminal, got: {:?}",
            agent.status
        );
        drop(agents);

        // A Done event should have been emitted.
        let event = rx.try_recv().expect("should emit Done event");
        assert!(
            matches!(event, AgentEvent::Done { .. }),
            "expected Done event, got: {event:?}"
        );
    }

    #[tokio::test]
    async fn server_owned_completion_fails_on_dirty_worktree() {
        use std::fs;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);
        // Create an uncommitted file.
        fs::write(repo.join("dirty.txt"), "not committed").unwrap();

        let pool = AgentPool::new(3001);
        pool.inject_test_agent_with_path(
            "s12",
            "coder-1",
            AgentStatus::Running,
            repo.to_path_buf(),
        );

        run_server_owned_completion(&pool.agents, pool.port, "s12", "coder-1", None)
            .await;

        let agents = pool.agents.lock().unwrap();
        let key = composite_key("s12", "coder-1");
        let agent = agents.get(&key).unwrap();
        assert!(agent.completion.is_some());
        assert!(!agent.completion.as_ref().unwrap().gates_passed);
        assert_eq!(agent.status, AgentStatus::Failed);
        assert!(
            agent
                .completion
                .as_ref()
                .unwrap()
                .gate_output
                .contains("uncommitted"),
            "gate_output should mention uncommitted changes"
        );
    }

    #[tokio::test]
    async fn server_owned_completion_nonexistent_agent_is_noop() {
        let pool = AgentPool::new(3001);
        // Should not panic or error — just silently return.
        run_server_owned_completion(&pool.agents, pool.port, "nonexistent", "bot", None)
            .await;
    }

    // ── move_story_to_current tests ────────────────────────────────────────────
    // No git repo needed: the watcher handles commits asynchronously.

    fn init_git_repo(repo: &std::path::Path) {
        Command::new("git")
            .args(["init"])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["config", "user.email", "test@test.com"])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["config", "user.name", "Test"])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "--allow-empty", "-m", "init"])
            .current_dir(repo)
            .output()
            .unwrap();
    }

    #[test]
    fn move_story_to_current_moves_file() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let upcoming = root.join(".story_kit/work/1_upcoming");
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&upcoming).unwrap();
        fs::create_dir_all(&current).unwrap();
        fs::write(upcoming.join("10_story_foo.md"), "test").unwrap();

        move_story_to_current(root, "10_story_foo").unwrap();

        assert!(!upcoming.join("10_story_foo.md").exists());
        assert!(current.join("10_story_foo.md").exists());
    }

    #[test]
    fn move_story_to_current_is_idempotent_when_already_current() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("11_story_foo.md"), "test").unwrap();

        move_story_to_current(root, "11_story_foo").unwrap();
        assert!(current.join("11_story_foo.md").exists());
    }

    #[test]
    fn move_story_to_current_noop_when_not_in_upcoming() {
        let tmp = tempfile::tempdir().unwrap();
        assert!(move_story_to_current(tmp.path(), "99_missing").is_ok());
    }

    #[test]
    fn move_bug_to_current_moves_from_upcoming() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let upcoming = root.join(".story_kit/work/1_upcoming");
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&upcoming).unwrap();
        fs::create_dir_all(&current).unwrap();
        fs::write(upcoming.join("1_bug_test.md"), "# Bug 1\n").unwrap();

        move_story_to_current(root, "1_bug_test").unwrap();

        assert!(!upcoming.join("1_bug_test.md").exists());
        assert!(current.join("1_bug_test.md").exists());
    }

    #[test]
    fn close_bug_moves_from_current_to_archive() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("2_bug_test.md"), "# Bug 2\n").unwrap();

        close_bug_to_archive(root, "2_bug_test").unwrap();

        assert!(!current.join("2_bug_test.md").exists());
        assert!(root.join(".story_kit/work/5_done/2_bug_test.md").exists());
    }

    #[test]
    fn close_bug_moves_from_upcoming_when_not_started() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let upcoming = root.join(".story_kit/work/1_upcoming");
        fs::create_dir_all(&upcoming).unwrap();
        fs::write(upcoming.join("3_bug_test.md"), "# Bug 3\n").unwrap();

        close_bug_to_archive(root, "3_bug_test").unwrap();

        assert!(!upcoming.join("3_bug_test.md").exists());
        assert!(root.join(".story_kit/work/5_done/3_bug_test.md").exists());
    }

    #[test]
    fn item_type_from_id_detects_types() {
        assert_eq!(item_type_from_id("1_bug_test"), "bug");
        assert_eq!(item_type_from_id("1_spike_research"), "spike");
        assert_eq!(item_type_from_id("50_story_my_story"), "story");
        assert_eq!(item_type_from_id("1_story_simple"), "story");
    }

    // ── pipeline_stage tests ──────────────────────────────────────────────────

    #[test]
    fn pipeline_stage_detects_coders() {
        assert_eq!(pipeline_stage("coder-1"), PipelineStage::Coder);
        assert_eq!(pipeline_stage("coder-2"), PipelineStage::Coder);
        assert_eq!(pipeline_stage("coder-3"), PipelineStage::Coder);
    }

    #[test]
    fn pipeline_stage_detects_qa() {
        assert_eq!(pipeline_stage("qa"), PipelineStage::Qa);
    }

    #[test]
    fn pipeline_stage_detects_mergemaster() {
        assert_eq!(pipeline_stage("mergemaster"), PipelineStage::Mergemaster);
    }

    #[test]
    fn pipeline_stage_supervisor_is_other() {
        assert_eq!(pipeline_stage("supervisor"), PipelineStage::Other);
        assert_eq!(pipeline_stage("default"), PipelineStage::Other);
        assert_eq!(pipeline_stage("unknown"), PipelineStage::Other);
    }

    // ── pipeline advance tests ────────────────────────────────────────────────

    #[tokio::test]
    async fn pipeline_advance_coder_gates_pass_moves_story_to_qa() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        // Set up story in 2_current/
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("50_story_test.md"), "test").unwrap();

        let pool = AgentPool::new(3001);
        pool.inject_test_agent_with_completion(
            "50_story_test",
            "coder-1",
            AgentStatus::Completed,
            root.to_path_buf(),
            CompletionReport {
                summary: "done".to_string(),
                gates_passed: true,
                gate_output: String::new(),
            },
        );

        // Call pipeline advance directly (bypasses background spawn for testing).
        pool.run_pipeline_advance_for_completed_agent("50_story_test", "coder-1")
            .await;

        // Story should have moved to 3_qa/ (start_agent for qa will fail in tests but
        // the file move happens before that).
        assert!(
            root.join(".story_kit/work/3_qa/50_story_test.md").exists(),
            "story should be in 3_qa/"
        );
        assert!(
            !current.join("50_story_test.md").exists(),
            "story should not still be in 2_current/"
        );
    }

    #[tokio::test]
    async fn pipeline_advance_qa_gates_pass_moves_story_to_merge() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        // Set up story in 3_qa/
        let qa_dir = root.join(".story_kit/work/3_qa");
        fs::create_dir_all(&qa_dir).unwrap();
        fs::write(qa_dir.join("51_story_test.md"), "test").unwrap();

        let pool = AgentPool::new(3001);
        pool.inject_test_agent_with_completion(
            "51_story_test",
            "qa",
            AgentStatus::Completed,
            root.to_path_buf(),
            CompletionReport {
                summary: "QA done".to_string(),
                gates_passed: true,
                gate_output: String::new(),
            },
        );

        pool.run_pipeline_advance_for_completed_agent("51_story_test", "qa")
            .await;

        // Story should have moved to 4_merge/
        assert!(
            root.join(".story_kit/work/4_merge/51_story_test.md").exists(),
            "story should be in 4_merge/"
        );
        assert!(
            !qa_dir.join("51_story_test.md").exists(),
            "story should not still be in 3_qa/"
        );
    }

    #[tokio::test]
    async fn pipeline_advance_supervisor_does_not_advance() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("52_story_test.md"), "test").unwrap();

        let pool = AgentPool::new(3001);
        pool.inject_test_agent_with_completion(
            "52_story_test",
            "supervisor",
            AgentStatus::Completed,
            root.to_path_buf(),
            CompletionReport {
                summary: "supervised".to_string(),
                gates_passed: true,
                gate_output: String::new(),
            },
        );

        pool.run_pipeline_advance_for_completed_agent("52_story_test", "supervisor")
            .await;

        // Story should NOT have moved (supervisors don't advance pipeline)
        assert!(
            current.join("52_story_test.md").exists(),
            "story should still be in 2_current/ for supervisor"
        );
    }

    // ── move_story_to_merge tests ──────────────────────────────────────────────

    #[test]
    fn move_story_to_merge_moves_file() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("20_story_foo.md"), "test").unwrap();

        move_story_to_merge(root, "20_story_foo").unwrap();

        assert!(!current.join("20_story_foo.md").exists());
        assert!(root.join(".story_kit/work/4_merge/20_story_foo.md").exists());
    }

    #[test]
    fn move_story_to_merge_from_qa_dir() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let qa_dir = root.join(".story_kit/work/3_qa");
        fs::create_dir_all(&qa_dir).unwrap();
        fs::write(qa_dir.join("40_story_test.md"), "test").unwrap();

        move_story_to_merge(root, "40_story_test").unwrap();

        assert!(!qa_dir.join("40_story_test.md").exists());
        assert!(root.join(".story_kit/work/4_merge/40_story_test.md").exists());
    }

    #[test]
    fn move_story_to_merge_idempotent_when_already_in_merge() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let merge_dir = root.join(".story_kit/work/4_merge");
        fs::create_dir_all(&merge_dir).unwrap();
        fs::write(merge_dir.join("21_story_test.md"), "test").unwrap();

        move_story_to_merge(root, "21_story_test").unwrap();
        assert!(merge_dir.join("21_story_test.md").exists());
    }

    #[test]
    fn move_story_to_merge_errors_when_not_in_current_or_qa() {
        let tmp = tempfile::tempdir().unwrap();
        let result = move_story_to_merge(tmp.path(), "99_nonexistent");
        assert!(result.unwrap_err().contains("not found in work/2_current/ or work/3_qa/"));
    }

    // ── move_story_to_qa tests ────────────────────────────────────────────────

    #[test]
    fn move_story_to_qa_moves_file() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("30_story_qa.md"), "test").unwrap();

        move_story_to_qa(root, "30_story_qa").unwrap();

        assert!(!current.join("30_story_qa.md").exists());
        assert!(root.join(".story_kit/work/3_qa/30_story_qa.md").exists());
    }

    #[test]
    fn move_story_to_qa_idempotent_when_already_in_qa() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let qa_dir = root.join(".story_kit/work/3_qa");
        fs::create_dir_all(&qa_dir).unwrap();
        fs::write(qa_dir.join("31_story_test.md"), "test").unwrap();

        move_story_to_qa(root, "31_story_test").unwrap();
        assert!(qa_dir.join("31_story_test.md").exists());
    }

    #[test]
    fn move_story_to_qa_errors_when_not_in_current() {
        let tmp = tempfile::tempdir().unwrap();
        let result = move_story_to_qa(tmp.path(), "99_nonexistent");
        assert!(result.unwrap_err().contains("not found in work/2_current/"));
    }

    // ── move_story_to_archived tests ──────────────────────────────────────────

    #[test]
    fn move_story_to_archived_finds_in_merge_dir() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let merge_dir = root.join(".story_kit/work/4_merge");
        fs::create_dir_all(&merge_dir).unwrap();
        fs::write(merge_dir.join("22_story_test.md"), "test").unwrap();

        move_story_to_archived(root, "22_story_test").unwrap();

        assert!(!merge_dir.join("22_story_test.md").exists());
        assert!(root.join(".story_kit/work/5_done/22_story_test.md").exists());
    }

    #[test]
    fn move_story_to_archived_error_when_not_in_current_or_merge() {
        let tmp = tempfile::tempdir().unwrap();
        let result = move_story_to_archived(tmp.path(), "99_nonexistent");
        assert!(result.unwrap_err().contains("4_merge"));
    }

    // ── merge_agent_work tests ────────────────────────────────────────────────

    #[tokio::test]
    async fn merge_agent_work_returns_error_when_branch_not_found() {
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        let pool = AgentPool::new(3001);
        // branch feature/story-99_nonexistent does not exist
        let result = pool
            .merge_agent_work(repo, "99_nonexistent")
            .await
            .unwrap();
        // Should fail (no branch) — not panic
        assert!(!result.success, "should fail when branch missing");
    }

    #[tokio::test]
    async fn merge_agent_work_succeeds_on_clean_branch() {
        use std::fs;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        // Create a feature branch with a commit
        Command::new("git")
            .args(["checkout", "-b", "feature/story-23_test"])
            .current_dir(repo)
            .output()
            .unwrap();
        fs::write(repo.join("feature.txt"), "feature content").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "add feature"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Switch back to master (initial branch)
        Command::new("git")
            .args(["checkout", "master"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Create the story file in 4_merge/ so we can test archival
        let merge_dir = repo.join(".story_kit/work/4_merge");
        fs::create_dir_all(&merge_dir).unwrap();
        let story_file = merge_dir.join("23_test.md");
        fs::write(&story_file, "---\nname: Test\n---\n").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "add story in merge"])
            .current_dir(repo)
            .output()
            .unwrap();

        let pool = AgentPool::new(3001);
        let report = pool.merge_agent_work(repo, "23_test").await.unwrap();

        // Merge should succeed (gates will run but cargo/pnpm results will depend on env)
        // At minimum the merge itself should succeed
        assert!(!report.had_conflicts, "should have no conflicts");
        // Note: gates_passed may be false in test env without Rust project, that's OK
        // The important thing is the merge itself ran
        assert!(
            report.success || report.gate_output.contains("Failed to run") || !report.gates_passed,
            "report should be coherent: {report:?}"
        );
        // Story should be in done if gates passed
        if report.story_archived {
            let done = repo.join(".story_kit/work/5_done/23_test.md");
            assert!(done.exists(), "done file should exist");
        }
    }

    // ── quality gate ordering test ────────────────────────────────

    /// Regression test for bug 142: quality gates must run BEFORE the fast-forward
    /// to master so that broken code never lands on master.
    ///
    /// Setup: a repo with a failing `script/test`, a feature branch with one commit.
    /// When `run_squash_merge` is called, the gates must detect failure and abort the
    /// fast-forward, leaving master HEAD unchanged.
    #[cfg(unix)]
    #[test]
    fn quality_gates_run_before_fast_forward_to_master() {
        use std::fs;
        use std::os::unix::fs::PermissionsExt;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        // Add a failing script/test so quality gates will fail.
        let script_dir = repo.join("script");
        fs::create_dir_all(&script_dir).unwrap();
        let script_test = script_dir.join("test");
        fs::write(&script_test, "#!/usr/bin/env bash\nexit 1\n").unwrap();
        let mut perms = fs::metadata(&script_test).unwrap().permissions();
        perms.set_mode(0o755);
        fs::set_permissions(&script_test, perms).unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "add failing script/test"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Create a feature branch with a commit.
        Command::new("git")
            .args(["checkout", "-b", "feature/story-142_test"])
            .current_dir(repo)
            .output()
            .unwrap();
        fs::write(repo.join("change.txt"), "feature change").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "feature work"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Switch back to master and record its HEAD.
        Command::new("git")
            .args(["checkout", "master"])
            .current_dir(repo)
            .output()
            .unwrap();
        let head_before = String::from_utf8(
            Command::new("git")
                .args(["rev-parse", "HEAD"])
                .current_dir(repo)
                .output()
                .unwrap()
                .stdout,
        )
        .unwrap()
        .trim()
        .to_string();

        // Run the squash-merge.  The failing script/test makes quality gates
        // fail → fast-forward must NOT happen.
        let result = run_squash_merge(repo, "feature/story-142_test", "142_test").unwrap();

        let head_after = String::from_utf8(
            Command::new("git")
                .args(["rev-parse", "HEAD"])
                .current_dir(repo)
                .output()
                .unwrap()
                .stdout,
        )
        .unwrap()
        .trim()
        .to_string();

        // Gates must have failed (script/test exits 1) so master should be untouched.
        assert!(
            !result.success,
            "run_squash_merge must report failure when gates fail"
        );
        assert_eq!(
            head_before, head_after,
            "master HEAD must not advance when quality gates fail (bug 142)"
        );
    }

    // ── run_project_tests tests ───────────────────────────────────

    #[cfg(unix)]
    #[test]
    fn run_project_tests_uses_script_test_when_present_and_passes() {
        use std::fs;
        use std::os::unix::fs::PermissionsExt;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let path = tmp.path();
        let script_dir = path.join("script");
        fs::create_dir_all(&script_dir).unwrap();
        let script_test = script_dir.join("test");
        fs::write(&script_test, "#!/usr/bin/env bash\necho 'all tests passed'\nexit 0\n").unwrap();
        let mut perms = fs::metadata(&script_test).unwrap().permissions();
        perms.set_mode(0o755);
        fs::set_permissions(&script_test, perms).unwrap();

        let (passed, output) = run_project_tests(path).unwrap();
        assert!(passed, "script/test exiting 0 should pass");
        assert!(output.contains("script/test"), "output should mention script/test");
    }

    #[cfg(unix)]
    #[test]
    fn run_project_tests_reports_failure_when_script_test_exits_nonzero() {
        use std::fs;
        use std::os::unix::fs::PermissionsExt;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let path = tmp.path();
        let script_dir = path.join("script");
        fs::create_dir_all(&script_dir).unwrap();
        let script_test = script_dir.join("test");
        fs::write(&script_test, "#!/usr/bin/env bash\nexit 1\n").unwrap();
        let mut perms = fs::metadata(&script_test).unwrap().permissions();
        perms.set_mode(0o755);
        fs::set_permissions(&script_test, perms).unwrap();

        let (passed, output) = run_project_tests(path).unwrap();
        assert!(!passed, "script/test exiting 1 should fail");
        assert!(output.contains("script/test"), "output should mention script/test");
    }

    // ── run_coverage_gate tests ───────────────────────────────────────────────

    #[cfg(unix)]
    #[test]
    fn coverage_gate_passes_when_script_absent() {
        use tempfile::tempdir;
        let tmp = tempdir().unwrap();
        let (passed, output) = run_coverage_gate(tmp.path()).unwrap();
        assert!(passed, "coverage gate should pass when script is absent");
        assert!(
            output.contains("not found"),
            "output should mention script not found"
        );
    }

    #[cfg(unix)]
    #[test]
    fn coverage_gate_passes_when_script_exits_zero() {
        use std::fs;
        use std::os::unix::fs::PermissionsExt;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let path = tmp.path();
        let script_dir = path.join("script");
        fs::create_dir_all(&script_dir).unwrap();
        let script = script_dir.join("test_coverage");
        fs::write(
            &script,
            "#!/usr/bin/env bash\necho 'Rust line coverage: 85%'\necho 'PASS: Coverage 85% meets threshold 0%'\nexit 0\n",
        )
        .unwrap();
        let mut perms = fs::metadata(&script).unwrap().permissions();
        perms.set_mode(0o755);
        fs::set_permissions(&script, perms).unwrap();

        let (passed, output) = run_coverage_gate(path).unwrap();
        assert!(passed, "coverage gate should pass when script exits 0");
        assert!(
            output.contains("script/test_coverage"),
            "output should mention script/test_coverage"
        );
    }

    #[cfg(unix)]
    #[test]
    fn coverage_gate_fails_when_script_exits_nonzero() {
        use std::fs;
        use std::os::unix::fs::PermissionsExt;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let path = tmp.path();
        let script_dir = path.join("script");
        fs::create_dir_all(&script_dir).unwrap();
        let script = script_dir.join("test_coverage");
        fs::write(
            &script,
            "#!/usr/bin/env bash\necho 'FAIL: Coverage 40% is below threshold 80%'\nexit 1\n",
        )
        .unwrap();
        let mut perms = fs::metadata(&script).unwrap().permissions();
        perms.set_mode(0o755);
        fs::set_permissions(&script, perms).unwrap();

        let (passed, output) = run_coverage_gate(path).unwrap();
        assert!(!passed, "coverage gate should fail when script exits 1");
        assert!(
            output.contains("script/test_coverage"),
            "output should mention script/test_coverage"
        );
    }

    // ── auto-assign helper tests ───────────────────────────────────

    #[test]
    fn scan_stage_items_returns_empty_for_missing_dir() {
        let tmp = tempfile::tempdir().unwrap();
        let items = scan_stage_items(tmp.path(), "2_current");
        assert!(items.is_empty());
    }

    #[test]
    fn scan_stage_items_returns_sorted_story_ids() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let stage_dir = tmp.path().join(".story_kit").join("work").join("2_current");
        fs::create_dir_all(&stage_dir).unwrap();
        fs::write(stage_dir.join("42_story_foo.md"), "---\nname: foo\n---").unwrap();
        fs::write(stage_dir.join("10_story_bar.md"), "---\nname: bar\n---").unwrap();
        fs::write(stage_dir.join("5_story_baz.md"), "---\nname: baz\n---").unwrap();
        // non-md file should be ignored
        fs::write(stage_dir.join("README.txt"), "ignore me").unwrap();

        let items = scan_stage_items(tmp.path(), "2_current");
        assert_eq!(items, vec!["10_story_bar", "42_story_foo", "5_story_baz"]);
    }

    #[test]
    fn is_story_assigned_returns_true_for_running_coder() {
        use crate::config::ProjectConfig;
        let config = ProjectConfig::default();
        let pool = AgentPool::new(3001);
        pool.inject_test_agent("42_story_foo", "coder-1", AgentStatus::Running);

        let agents = pool.agents.lock().unwrap();
        assert!(is_story_assigned_for_stage(
            &config,
            &agents,
            "42_story_foo",
            &PipelineStage::Coder
        ));
        // Same story but wrong stage — should be false
        assert!(!is_story_assigned_for_stage(
            &config,
            &agents,
            "42_story_foo",
            &PipelineStage::Qa
        ));
        // Different story — should be false
        assert!(!is_story_assigned_for_stage(
            &config,
            &agents,
            "99_story_other",
            &PipelineStage::Coder
        ));
    }

    #[test]
    fn is_story_assigned_returns_false_for_completed_agent() {
        use crate::config::ProjectConfig;
        let config = ProjectConfig::default();
        let pool = AgentPool::new(3001);
        pool.inject_test_agent("42_story_foo", "coder-1", AgentStatus::Completed);

        let agents = pool.agents.lock().unwrap();
        // Completed agents don't count as assigned
        assert!(!is_story_assigned_for_stage(
            &config,
            &agents,
            "42_story_foo",
            &PipelineStage::Coder
        ));
    }

    #[test]
    fn find_free_agent_returns_none_when_all_busy() {
        use crate::config::ProjectConfig;
        let config = ProjectConfig::parse(
            r#"
[[agent]]
name = "coder-1"
[[agent]]
name = "coder-2"
"#,
        )
        .unwrap();

        let pool = AgentPool::new(3001);
        pool.inject_test_agent("s1", "coder-1", AgentStatus::Running);
        pool.inject_test_agent("s2", "coder-2", AgentStatus::Running);

        let agents = pool.agents.lock().unwrap();
        let free = find_free_agent_for_stage(&config, &agents, &PipelineStage::Coder);
        assert!(free.is_none(), "no free coders should be available");
    }

    #[test]
    fn find_free_agent_returns_first_free_coder() {
        use crate::config::ProjectConfig;
        let config = ProjectConfig::parse(
            r#"
[[agent]]
name = "coder-1"
[[agent]]
name = "coder-2"
[[agent]]
name = "coder-3"
"#,
        )
        .unwrap();

        let pool = AgentPool::new(3001);
        // coder-1 is busy, coder-2 is free
        pool.inject_test_agent("s1", "coder-1", AgentStatus::Running);

        let agents = pool.agents.lock().unwrap();
        let free = find_free_agent_for_stage(&config, &agents, &PipelineStage::Coder);
        assert_eq!(free, Some("coder-2"), "coder-2 should be the first free coder");
    }

    #[test]
    fn find_free_agent_ignores_completed_agents() {
        use crate::config::ProjectConfig;
        let config = ProjectConfig::parse(
            r#"
[[agent]]
name = "coder-1"
"#,
        )
        .unwrap();

        let pool = AgentPool::new(3001);
        // coder-1 completed its previous story — it's free for a new one
        pool.inject_test_agent("s1", "coder-1", AgentStatus::Completed);

        let agents = pool.agents.lock().unwrap();
        let free = find_free_agent_for_stage(&config, &agents, &PipelineStage::Coder);
        assert_eq!(free, Some("coder-1"), "completed coder-1 should be free");
    }

    #[test]
    fn find_free_agent_returns_none_for_wrong_stage() {
        use crate::config::ProjectConfig;
        let config = ProjectConfig::parse(
            r#"
[[agent]]
name = "qa"
"#,
        )
        .unwrap();

        let agents: HashMap<String, StoryAgent> = HashMap::new();
        // Looking for a Coder but only QA is configured
        let free = find_free_agent_for_stage(&config, &agents, &PipelineStage::Coder);
        assert!(free.is_none());
        // Looking for QA should find it
        let free_qa = find_free_agent_for_stage(&config, &agents, &PipelineStage::Qa);
        assert_eq!(free_qa, Some("qa"));
    }

    // ── agent_config_stage / stage field tests ────────────────────────────────

    #[test]
    fn find_free_agent_uses_config_stage_field_not_name() {
        // Agents named "qa-2" and "coder-opus" don't match the legacy name heuristic
        // but should be picked up via their explicit stage field.
        use crate::config::ProjectConfig;
        let config = ProjectConfig::parse(
            r#"
[[agent]]
name = "qa-2"
stage = "qa"

[[agent]]
name = "coder-opus"
stage = "coder"
"#,
        )
        .unwrap();

        let agents: HashMap<String, StoryAgent> = HashMap::new();

        // qa-2 should be found for PipelineStage::Qa via config stage field
        let free_qa = find_free_agent_for_stage(&config, &agents, &PipelineStage::Qa);
        assert_eq!(free_qa, Some("qa-2"), "qa-2 with stage=qa should be found");

        // coder-opus should be found for PipelineStage::Coder via config stage field
        let free_coder = find_free_agent_for_stage(&config, &agents, &PipelineStage::Coder);
        assert_eq!(
            free_coder,
            Some("coder-opus"),
            "coder-opus with stage=coder should be found"
        );

        // Neither should match the other stage
        let free_merge = find_free_agent_for_stage(&config, &agents, &PipelineStage::Mergemaster);
        assert!(free_merge.is_none());
    }

    #[test]
    fn is_story_assigned_uses_config_stage_field_for_nonstandard_names() {
        use crate::config::ProjectConfig;
        let config = ProjectConfig::parse(
            r#"
[[agent]]
name = "qa-2"
stage = "qa"
"#,
        )
        .unwrap();

        let pool = AgentPool::new(3001);
        pool.inject_test_agent("42_story_foo", "qa-2", AgentStatus::Running);

        let agents = pool.agents.lock().unwrap();
        // qa-2 with stage=qa should be recognised as a QA agent
        assert!(
            is_story_assigned_for_stage(&config, &agents, "42_story_foo", &PipelineStage::Qa),
            "qa-2 should be detected as assigned to QA stage"
        );
        // Should NOT appear as a coder
        assert!(
            !is_story_assigned_for_stage(
                &config,
                &agents,
                "42_story_foo",
                &PipelineStage::Coder
            ),
            "qa-2 should not be detected as a coder"
        );
    }

    // ── find_active_story_stage tests ─────────────────────────────────────────

    #[test]
    fn find_active_story_stage_detects_current() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("10_story_test.md"), "test").unwrap();

        assert_eq!(
            find_active_story_stage(root, "10_story_test"),
            Some("2_current")
        );
    }

    #[test]
    fn find_active_story_stage_detects_qa() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let qa = root.join(".story_kit/work/3_qa");
        fs::create_dir_all(&qa).unwrap();
        fs::write(qa.join("11_story_test.md"), "test").unwrap();

        assert_eq!(
            find_active_story_stage(root, "11_story_test"),
            Some("3_qa")
        );
    }

    #[test]
    fn find_active_story_stage_detects_merge() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();
        let merge = root.join(".story_kit/work/4_merge");
        fs::create_dir_all(&merge).unwrap();
        fs::write(merge.join("12_story_test.md"), "test").unwrap();

        assert_eq!(
            find_active_story_stage(root, "12_story_test"),
            Some("4_merge")
        );
    }

    #[test]
    fn find_active_story_stage_returns_none_for_unknown_story() {
        let tmp = tempfile::tempdir().unwrap();
        assert_eq!(find_active_story_stage(tmp.path(), "99_nonexistent"), None);
    }

    // ── start_agent single-instance concurrency tests ─────────────────────────

    /// Regression test for bug 97: the agent pool must reject a second concurrent
    /// instance of the same agent name even if it would run on a different story.
    #[tokio::test]
    async fn start_agent_rejects_when_same_agent_already_running_on_another_story() {
        use std::fs;

        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        // Write a minimal project.toml so ProjectConfig::load can find the "qa" agent.
        let sk_dir = root.join(".story_kit");
        fs::create_dir_all(&sk_dir).unwrap();
        fs::write(
            sk_dir.join("project.toml"),
            "[[agent]]\nname = \"qa\"\n",
        )
        .unwrap();

        let pool = AgentPool::new(3001);
        // Simulate qa already running on story-a.
        pool.inject_test_agent("story-a", "qa", AgentStatus::Running);

        // Attempt to start qa on story-b — must be rejected.
        let result = pool
            .start_agent(root, "story-b", Some("qa"), None)
            .await;

        assert!(
            result.is_err(),
            "start_agent should fail when qa is already running on another story"
        );
        let err = result.unwrap_err();
        assert!(
            err.contains("already running") || err.contains("becomes available"),
            "error message should explain why: got '{err}'"
        );
    }

    /// Verify that the concurrency guard does NOT block an agent that is merely
    /// Completed (not Running/Pending) — completed agents are free for new work.
    #[tokio::test]
    async fn start_agent_allows_new_story_when_previous_run_is_completed() {
        use std::fs;

        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        let sk_dir = root.join(".story_kit");
        fs::create_dir_all(&sk_dir).unwrap();
        fs::write(
            sk_dir.join("project.toml"),
            "[[agent]]\nname = \"qa\"\n",
        )
        .unwrap();

        let pool = AgentPool::new(3001);
        // Previous run completed — should NOT block a new story.
        pool.inject_test_agent("story-a", "qa", AgentStatus::Completed);

        // The call will fail eventually (no real worktree / Claude CLI), but it must
        // NOT fail at the concurrency check.  We detect the difference by inspecting
        // the error message: a concurrency rejection says "already running", while a
        // later failure (missing story file, missing claude binary, etc.) says something else.
        let result = pool
            .start_agent(root, "story-b", Some("qa"), None)
            .await;

        if let Err(ref e) = result {
            assert!(
                !e.contains("already running") && !e.contains("becomes available"),
                "completed agent must not trigger the concurrency guard: got '{e}'"
            );
        }
        // result may be Ok (unlikely in test env) or Err for infra reasons — both fine.
    }

    // ── worktree_has_committed_work tests ─────────────────────────────────────

    #[test]
    fn worktree_has_committed_work_false_on_fresh_repo() {
        let tmp = tempfile::tempdir().unwrap();
        let repo = tmp.path();
        // init_git_repo creates the initial commit on the default branch.
        // HEAD IS the base branch — no commits ahead.
        init_git_repo(repo);
        assert!(!worktree_has_committed_work(repo));
    }

    #[test]
    fn worktree_has_committed_work_true_after_commit_on_feature_branch() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let project_root = tmp.path().join("project");
        fs::create_dir_all(&project_root).unwrap();
        init_git_repo(&project_root);

        // Create a git worktree on a feature branch.
        let wt_path = tmp.path().join("wt");
        Command::new("git")
            .args([
                "worktree",
                "add",
                &wt_path.to_string_lossy(),
                "-b",
                "feature/story-99_test",
            ])
            .current_dir(&project_root)
            .output()
            .unwrap();

        // No commits on the feature branch yet — same as base branch.
        assert!(!worktree_has_committed_work(&wt_path));

        // Add a commit to the feature branch in the worktree.
        fs::write(wt_path.join("work.txt"), "done").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(&wt_path)
            .output()
            .unwrap();
        Command::new("git")
            .args([
                "-c",
                "user.email=test@test.com",
                "-c",
                "user.name=Test",
                "commit",
                "-m",
                "coder: implement story",
            ])
            .current_dir(&wt_path)
            .output()
            .unwrap();

        // Now the feature branch is ahead of the base branch.
        assert!(worktree_has_committed_work(&wt_path));
    }

    // ── reconcile_on_startup tests ────────────────────────────────────────────

    #[tokio::test]
    async fn reconcile_on_startup_noop_when_no_worktrees() {
        let tmp = tempfile::tempdir().unwrap();
        let pool = AgentPool::new(3001);
        let (tx, _rx) = broadcast::channel(16);
        // Should not panic; no worktrees to reconcile.
        pool.reconcile_on_startup(tmp.path(), &tx).await;
    }

    #[tokio::test]
    async fn reconcile_on_startup_emits_done_event() {
        let tmp = tempfile::tempdir().unwrap();
        let pool = AgentPool::new(3001);
        let (tx, mut rx) = broadcast::channel::<ReconciliationEvent>(16);
        pool.reconcile_on_startup(tmp.path(), &tx).await;

        // Collect all events; the last must be "done".
        let mut events: Vec<ReconciliationEvent> = Vec::new();
        while let Ok(evt) = rx.try_recv() {
            events.push(evt);
        }
        assert!(
            events.iter().any(|e| e.status == "done"),
            "reconcile_on_startup must emit a 'done' event; got: {:?}",
            events.iter().map(|e| &e.status).collect::<Vec<_>>()
        );
    }

    #[tokio::test]
    async fn reconcile_on_startup_skips_story_without_committed_work() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        // Set up story in 2_current/.
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("60_story_test.md"), "test").unwrap();

        // Create a worktree directory that is a fresh git repo with no commits
        // ahead of its own base branch (simulates a worktree where no work was done).
        let wt_dir = root.join(".story_kit/worktrees/60_story_test");
        fs::create_dir_all(&wt_dir).unwrap();
        init_git_repo(&wt_dir);

        let pool = AgentPool::new(3001);
        let (tx, _rx) = broadcast::channel(16);
        pool.reconcile_on_startup(root, &tx).await;

        // Story should still be in 2_current/ — nothing was reconciled.
        assert!(
            current.join("60_story_test.md").exists(),
            "story should stay in 2_current/ when worktree has no committed work"
        );
    }

    #[tokio::test]
    async fn reconcile_on_startup_runs_gates_on_worktree_with_committed_work() {
        use std::fs;
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        // Set up a git repo for the project root.
        init_git_repo(root);

        // Set up story in 2_current/ and commit it so the project root is clean.
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("61_story_test.md"), "test").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(root)
            .output()
            .unwrap();
        Command::new("git")
            .args([
                "-c",
                "user.email=test@test.com",
                "-c",
                "user.name=Test",
                "commit",
                "-m",
                "add story",
            ])
            .current_dir(root)
            .output()
            .unwrap();

        // Create a real git worktree for the story.
        let wt_dir = root.join(".story_kit/worktrees/61_story_test");
        fs::create_dir_all(wt_dir.parent().unwrap()).unwrap();
        Command::new("git")
            .args([
                "worktree",
                "add",
                &wt_dir.to_string_lossy(),
                "-b",
                "feature/story-61_story_test",
            ])
            .current_dir(root)
            .output()
            .unwrap();

        // Add a commit to the feature branch (simulates coder completing work).
        fs::write(wt_dir.join("implementation.txt"), "done").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(&wt_dir)
            .output()
            .unwrap();
        Command::new("git")
            .args([
                "-c",
                "user.email=test@test.com",
                "-c",
                "user.name=Test",
                "commit",
                "-m",
                "implement story",
            ])
            .current_dir(&wt_dir)
            .output()
            .unwrap();

        assert!(
            worktree_has_committed_work(&wt_dir),
            "test setup: worktree should have committed work"
        );

        let pool = AgentPool::new(3001);
        let (tx, _rx) = broadcast::channel(16);
        pool.reconcile_on_startup(root, &tx).await;

        // In the test env, cargo clippy will fail (no Cargo.toml) so gates fail
        // and the story stays in 2_current/.  The important assertion is that
        // reconcile ran without panicking and the story is in a consistent state.
        let in_current = current.join("61_story_test.md").exists();
        let in_qa = root
            .join(".story_kit/work/3_qa/61_story_test.md")
            .exists();
        assert!(
            in_current || in_qa,
            "story should be in 2_current/ or 3_qa/ after reconciliation"
        );
    }

    #[test]
    fn test_emit_event_writes_to_log_writer() {
        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        let log_writer =
            AgentLogWriter::new(root, "42_story_foo", "coder-1", "sess-emit").unwrap();
        let log_mutex = Mutex::new(log_writer);

        let (tx, _rx) = broadcast::channel::<AgentEvent>(64);
        let event_log: Mutex<Vec<AgentEvent>> = Mutex::new(Vec::new());

        let event = AgentEvent::Status {
            story_id: "42_story_foo".to_string(),
            agent_name: "coder-1".to_string(),
            status: "running".to_string(),
        };

        emit_event(event, &tx, &event_log, Some(&log_mutex));

        // Verify event was added to in-memory log
        let mem_events = event_log.lock().unwrap();
        assert_eq!(mem_events.len(), 1);
        drop(mem_events);

        // Verify event was written to the log file
        let log_path =
            crate::agent_log::log_file_path(root, "42_story_foo", "coder-1", "sess-emit");
        let entries = crate::agent_log::read_log(&log_path).unwrap();
        assert_eq!(entries.len(), 1);
        assert_eq!(entries[0].event["type"], "status");
        assert_eq!(entries[0].event["status"], "running");
    }

    // ── bug 167: handle_agent_stream_event routes thinking/text correctly ───

    #[test]
    fn stream_event_thinking_delta_emits_thinking_event() {
        let (tx, mut rx) = broadcast::channel::<AgentEvent>(64);
        let event_log: Mutex<Vec<AgentEvent>> = Mutex::new(Vec::new());

        let event = serde_json::json!({
            "type": "content_block_delta",
            "delta": {"type": "thinking_delta", "thinking": "Let me analyze this..."}
        });

        handle_agent_stream_event(&event, "s1", "coder-1", &tx, &event_log, None);

        let received = rx.try_recv().unwrap();
        match received {
            AgentEvent::Thinking {
                story_id,
                agent_name,
                text,
            } => {
                assert_eq!(story_id, "s1");
                assert_eq!(agent_name, "coder-1");
                assert_eq!(text, "Let me analyze this...");
            }
            other => panic!("Expected Thinking event, got: {other:?}"),
        }
    }

    #[test]
    fn stream_event_text_delta_emits_output_event() {
        let (tx, mut rx) = broadcast::channel::<AgentEvent>(64);
        let event_log: Mutex<Vec<AgentEvent>> = Mutex::new(Vec::new());

        let event = serde_json::json!({
            "type": "content_block_delta",
            "delta": {"type": "text_delta", "text": "Here is the result."}
        });

        handle_agent_stream_event(&event, "s1", "coder-1", &tx, &event_log, None);

        let received = rx.try_recv().unwrap();
        match received {
            AgentEvent::Output {
                story_id,
                agent_name,
                text,
            } => {
                assert_eq!(story_id, "s1");
                assert_eq!(agent_name, "coder-1");
                assert_eq!(text, "Here is the result.");
            }
            other => panic!("Expected Output event, got: {other:?}"),
        }
    }

    #[test]
    fn stream_event_input_json_delta_ignored() {
        let (tx, mut rx) = broadcast::channel::<AgentEvent>(64);
        let event_log: Mutex<Vec<AgentEvent>> = Mutex::new(Vec::new());

        let event = serde_json::json!({
            "type": "content_block_delta",
            "delta": {"type": "input_json_delta", "partial_json": "{\"file\":"}
        });

        handle_agent_stream_event(&event, "s1", "coder-1", &tx, &event_log, None);

        // No event should be emitted for tool argument deltas
        assert!(rx.try_recv().is_err());
    }

    #[test]
    fn stream_event_non_delta_type_ignored() {
        let (tx, mut rx) = broadcast::channel::<AgentEvent>(64);
        let event_log: Mutex<Vec<AgentEvent>> = Mutex::new(Vec::new());

        let event = serde_json::json!({
            "type": "message_start",
            "message": {"role": "assistant"}
        });

        handle_agent_stream_event(&event, "s1", "coder-1", &tx, &event_log, None);

        assert!(rx.try_recv().is_err());
    }

    // ── bug 118: pending entry cleanup on start_agent failure ────────────────

    /// Regression test for bug 118: when worktree creation fails (e.g. because
    /// there is no git repo), the Pending entry that was inserted into the agent
    /// HashMap must not remain Pending — it must transition to Failed.  This
    /// prevents `find_free_agent_for_stage` / auto-assign from being permanently
    /// blocked.
    ///
    /// With story 157 the worktree creation moved into the background spawn, so
    /// `start_agent` returns `Ok(Pending)` immediately.  We use `wait_for_agent`
    /// to block until the background task resolves.
    #[tokio::test]
    async fn start_agent_cleans_up_pending_entry_on_failure() {
        use std::fs;

        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        // Minimal project.toml with a "qa" agent.
        let sk_dir = root.join(".story_kit");
        fs::create_dir_all(&sk_dir).unwrap();
        fs::write(
            sk_dir.join("project.toml"),
            "[[agent]]\nname = \"qa\"\n",
        )
        .unwrap();

        // Create the story in upcoming so `move_story_to_current` succeeds,
        // but do NOT init a git repo — `create_worktree` will fail in the spawn.
        let upcoming = root.join(".story_kit/work/1_upcoming");
        fs::create_dir_all(&upcoming).unwrap();
        fs::write(
            upcoming.join("50_story_test.md"),
            "---\nname: Test\n---\n",
        )
        .unwrap();

        let pool = AgentPool::new(3099);

        let result = pool
            .start_agent(root, "50_story_test", Some("qa"), None)
            .await;

        // With the non-blocking flow, start_agent returns Ok(Pending) immediately.
        // Worktree creation failure happens asynchronously in the background.
        assert!(
            result.is_ok(),
            "start_agent should return Ok(Pending) immediately: {:?}",
            result.err()
        );
        assert_eq!(
            result.unwrap().status,
            AgentStatus::Pending,
            "initial status must be Pending"
        );

        // Wait for the background task to reach a terminal state.
        // It must fail (no git repo → create_worktree returns an error).
        let final_info = pool
            .wait_for_agent("50_story_test", "qa", 5000)
            .await
            .expect("wait_for_agent should not time out");
        assert_eq!(
            final_info.status,
            AgentStatus::Failed,
            "agent must transition to Failed after worktree creation error"
        );

        // The pool must NOT retain a Pending or Running entry for this agent.
        let agents = pool.agents.lock().unwrap();
        let leaked = agents.values().any(|a| {
            a.agent_name == "qa"
                && matches!(a.status, AgentStatus::Pending | AgentStatus::Running)
        });
        assert!(
            !leaked,
            "agent pool must not retain a Pending/Running entry after worktree creation fails"
        );
    }

    /// Verify that a successful start_agent keeps the Running entry (guard is
    /// disarmed).  We cannot truly spawn an agent in tests, but we verify that
    /// the concurrency check still blocks a second concurrent start — which
    /// proves the first entry survived the guard.
    #[tokio::test]
    async fn start_agent_guard_does_not_remove_running_entry() {
        use std::fs;

        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        let sk_dir = root.join(".story_kit");
        fs::create_dir_all(&sk_dir).unwrap();
        fs::write(
            sk_dir.join("project.toml"),
            "[[agent]]\nname = \"qa\"\n",
        )
        .unwrap();

        let pool = AgentPool::new(3099);

        // Manually inject a Running agent (simulates successful start).
        pool.inject_test_agent("story-x", "qa", AgentStatus::Running);

        // Attempting to start the same agent on a different story must be
        // rejected — the Running entry must still be there.
        let result = pool
            .start_agent(root, "story-y", Some("qa"), None)
            .await;

        assert!(result.is_err());
        let err = result.unwrap_err();
        assert!(
            err.contains("already running") || err.contains("becomes available"),
            "running entry must survive: got '{err}'"
        );
    }

    // ── TOCTOU race-condition regression tests (story 132) ───────────────────

    /// Verify that a Pending entry (not just Running) blocks a concurrent
    /// start_agent for the same agent name on a different story.  This proves
    /// the check-and-insert is atomic: the Pending entry is visible to the
    /// second caller because it was inserted while the lock was still held.
    #[tokio::test]
    async fn toctou_pending_entry_blocks_same_agent_on_different_story() {
        use std::fs;

        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        let sk_dir = root.join(".story_kit");
        fs::create_dir_all(&sk_dir).unwrap();
        fs::write(sk_dir.join("project.toml"), "[[agent]]\nname = \"coder-1\"\n").unwrap();

        let pool = AgentPool::new(3099);

        // Simulate what the winning concurrent call would have done: insert a
        // Pending entry for coder-1 on story-86.
        pool.inject_test_agent("86_story_foo", "coder-1", AgentStatus::Pending);

        // Now attempt to start coder-1 on a *different* story — must be rejected.
        let result = pool
            .start_agent(root, "130_story_bar", Some("coder-1"), None)
            .await;

        assert!(result.is_err(), "second start_agent must be rejected");
        let err = result.unwrap_err();
        assert!(
            err.contains("already running") || err.contains("becomes available"),
            "expected concurrency-rejection message, got: '{err}'"
        );
    }

    /// Concurrent start_agent calls for the same agent name on different stories
    /// must result in exactly one rejection due to the concurrency check (not
    /// due to an unrelated failure such as missing git repo).
    #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
    async fn toctou_concurrent_start_agent_same_agent_exactly_one_concurrency_rejection() {
        use std::fs;
        use std::sync::Arc;

        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path().to_path_buf();

        let sk_dir = root.join(".story_kit");
        fs::create_dir_all(sk_dir.join("work/1_upcoming")).unwrap();
        fs::write(
            root.join(".story_kit/project.toml"),
            "[[agent]]\nname = \"coder-1\"\n",
        )
        .unwrap();
        // Both stories must exist in upcoming so move_story_to_current can run
        // (only the winner reaches that point, but we set both up defensively).
        fs::write(
            root.join(".story_kit/work/1_upcoming/86_story_foo.md"),
            "---\nname: Foo\n---\n",
        )
        .unwrap();
        fs::write(
            root.join(".story_kit/work/1_upcoming/130_story_bar.md"),
            "---\nname: Bar\n---\n",
        )
        .unwrap();

        let pool = Arc::new(AgentPool::new(3099));

        let pool1 = pool.clone();
        let root1 = root.clone();
        let t1 = tokio::spawn(async move {
            pool1
                .start_agent(&root1, "86_story_foo", Some("coder-1"), None)
                .await
        });

        let pool2 = pool.clone();
        let root2 = root.clone();
        let t2 = tokio::spawn(async move {
            pool2
                .start_agent(&root2, "130_story_bar", Some("coder-1"), None)
                .await
        });

        let (r1, r2) = tokio::join!(t1, t2);
        let r1 = r1.unwrap();
        let r2 = r2.unwrap();

        // The concurrency-rejection message always contains "already running" /
        // "becomes available".  Any other error (e.g., missing git repo) means
        // that call *won* the atomic check-and-insert.
        let concurrency_rejections = [&r1, &r2]
            .iter()
            .filter(|r| {
                r.as_ref().is_err_and(|e| {
                    e.contains("already running") || e.contains("becomes available")
                })
            })
            .count();

        assert_eq!(
            concurrency_rejections, 1,
            "exactly one call must be rejected by the concurrency check; \
             got r1={r1:?} r2={r2:?}"
        );
    }

    /// Two concurrent auto_assign_available_work calls must not assign the same
    /// agent to two stories simultaneously.  After both complete, at most one
    /// Pending/Running entry must exist per agent name.
    #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
    async fn toctou_concurrent_auto_assign_no_duplicate_agent_assignments() {
        use std::fs;
        use std::sync::Arc;

        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path().to_path_buf();

        let sk_dir = root.join(".story_kit");
        // Two stories waiting in 2_current, one coder agent.
        fs::create_dir_all(sk_dir.join("work/2_current")).unwrap();
        fs::write(
            sk_dir.join("project.toml"),
            "[[agent]]\nname = \"coder-1\"\n",
        )
        .unwrap();
        fs::write(
            sk_dir.join("work/2_current/86_story_foo.md"),
            "---\nname: Foo\n---\n",
        )
        .unwrap();
        fs::write(
            sk_dir.join("work/2_current/130_story_bar.md"),
            "---\nname: Bar\n---\n",
        )
        .unwrap();

        let pool = Arc::new(AgentPool::new(3099));

        // Run two concurrent auto_assign calls.
        let pool1 = pool.clone();
        let root1 = root.clone();
        let t1 = tokio::spawn(async move { pool1.auto_assign_available_work(&root1).await });

        let pool2 = pool.clone();
        let root2 = root.clone();
        let t2 = tokio::spawn(async move { pool2.auto_assign_available_work(&root2).await });

        let _ = tokio::join!(t1, t2);

        // At most one Pending/Running entry should exist for coder-1.
        let agents = pool.agents.lock().unwrap();
        let active_coder_count = agents
            .values()
            .filter(|a| {
                a.agent_name == "coder-1"
                    && matches!(a.status, AgentStatus::Pending | AgentStatus::Running)
            })
            .count();

        assert!(
            active_coder_count <= 1,
            "coder-1 must not be assigned to more than one story simultaneously; \
             found {active_coder_count} active entries"
        );
    }

    // ── resolve_simple_conflicts unit tests ──────────────────────────────────

    #[test]
    fn resolve_simple_conflicts_no_markers() {
        let input = "line 1\nline 2\nline 3\n";
        let result = resolve_simple_conflicts(input);
        assert_eq!(result, Some(input.to_string()));
    }

    #[test]
    fn resolve_simple_conflicts_additive() {
        let input = "\
before
ours line 1
ours line 2
theirs line 1
theirs line 2
after
";
        let result = resolve_simple_conflicts(input).unwrap();
        assert!(
            !result.contains("<<<<<<<"),
            "should not contain conflict markers"
        );
        assert!(
            !result.contains(">>>>>>>"),
            "should not contain conflict markers"
        );
        assert!(result.contains("ours line 1"));
        assert!(result.contains("ours line 2"));
        assert!(result.contains("theirs line 1"));
        assert!(result.contains("theirs line 2"));
        assert!(result.contains("before"));
        assert!(result.contains("after"));
        // Ours comes before theirs
        let ours_pos = result.find("ours line 1").unwrap();
        let theirs_pos = result.find("theirs line 1").unwrap();
        assert!(
            ours_pos < theirs_pos,
            "ours should come before theirs"
        );
    }

    #[test]
    fn resolve_simple_conflicts_multiple_blocks() {
        let input = "\
header
ours block 1
theirs block 1
middle
ours block 2
theirs block 2
footer
";
        let result = resolve_simple_conflicts(input).unwrap();
        assert!(!result.contains("<<<<<<<"));
        assert!(result.contains("ours block 1"));
        assert!(result.contains("theirs block 1"));
        assert!(result.contains("ours block 2"));
        assert!(result.contains("theirs block 2"));
        assert!(result.contains("header"));
        assert!(result.contains("middle"));
        assert!(result.contains("footer"));
    }

    #[test]
    fn resolve_simple_conflicts_malformed_no_separator() {
        let input = "\
<<<<<<< HEAD
ours
>>>>>>> feature
";
        let result = resolve_simple_conflicts(input);
        assert!(result.is_none(), "malformed conflict (no separator) should return None");
    }

    #[test]
    fn resolve_simple_conflicts_malformed_no_end() {
        let input = "\
<<<<<<< HEAD
ours
=======
theirs
";
        let result = resolve_simple_conflicts(input);
        assert!(result.is_none(), "malformed conflict (no end marker) should return None");
    }

    #[test]
    fn resolve_simple_conflicts_preserves_no_trailing_newline() {
        let input = "before\n<<<<<<< HEAD\nours\n=======\ntheirs\n>>>>>>> branch\nafter";
        let result = resolve_simple_conflicts(input).unwrap();
        assert!(!result.ends_with('\n'), "should not add trailing newline if original lacks one");
        assert!(result.ends_with("after"));
    }

    // ── merge-queue squash-merge integration tests ──────────────────────────

    #[tokio::test]
    async fn squash_merge_uses_merge_queue_no_conflict_markers_on_master() {
        use std::fs;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        // Create a file that will be conflicted on master.
        fs::write(repo.join("shared.txt"), "line 1\nline 2\n").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "initial shared file"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Create a feature branch that modifies the file.
        Command::new("git")
            .args(["checkout", "-b", "feature/story-conflict_test"])
            .current_dir(repo)
            .output()
            .unwrap();
        fs::write(repo.join("shared.txt"), "line 1\nline 2\nfeature addition\n").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "feature: add line"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Switch to master and make a conflicting change.
        Command::new("git")
            .args(["checkout", "master"])
            .current_dir(repo)
            .output()
            .unwrap();
        fs::write(repo.join("shared.txt"), "line 1\nline 2\nmaster addition\n").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "master: add line"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Run the squash merge.
        let result = run_squash_merge(repo, "feature/story-conflict_test", "conflict_test")
            .unwrap();

        // Master should NEVER contain conflict markers, regardless of outcome.
        let master_content = fs::read_to_string(repo.join("shared.txt")).unwrap();
        assert!(
            !master_content.contains("<<<<<<<"),
            "master must never contain conflict markers, got:\n{master_content}"
        );
        assert!(
            !master_content.contains(">>>>>>>"),
            "master must never contain conflict markers, got:\n{master_content}"
        );

        // The merge should have had conflicts.
        assert!(result.had_conflicts, "should detect conflicts");

        // Conflicts should have been auto-resolved (both are simple additions).
        if result.conflicts_resolved {
            assert!(result.success, "auto-resolved merge should succeed");
            assert!(
                master_content.contains("master addition"),
                "master side should be present"
            );
            assert!(
                master_content.contains("feature addition"),
                "feature side should be present"
            );
        }

        // Verify no leftover merge-queue branch.
        let branches = Command::new("git")
            .args(["branch", "--list", "merge-queue/*"])
            .current_dir(repo)
            .output()
            .unwrap();
        let branch_list = String::from_utf8_lossy(&branches.stdout);
        assert!(
            branch_list.trim().is_empty(),
            "merge-queue branch should be cleaned up, got: {branch_list}"
        );

        // Verify no leftover merge workspace directory.
        assert!(
            !repo.join(".story_kit/merge_workspace").exists(),
            "merge workspace should be cleaned up"
        );
    }

    #[tokio::test]
    async fn squash_merge_clean_merge_succeeds() {
        use std::fs;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        // Create feature branch with a new file.
        Command::new("git")
            .args(["checkout", "-b", "feature/story-clean_test"])
            .current_dir(repo)
            .output()
            .unwrap();
        fs::write(repo.join("new_file.txt"), "new content").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "add new file"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Switch back to master.
        Command::new("git")
            .args(["checkout", "master"])
            .current_dir(repo)
            .output()
            .unwrap();

        let result = run_squash_merge(repo, "feature/story-clean_test", "clean_test")
            .unwrap();

        assert!(result.success, "clean merge should succeed");
        assert!(!result.had_conflicts, "clean merge should have no conflicts");
        assert!(!result.conflicts_resolved, "no conflicts means nothing to resolve");
        assert!(
            repo.join("new_file.txt").exists(),
            "merged file should exist on master"
        );
    }

    #[tokio::test]
    async fn squash_merge_nonexistent_branch_fails() {
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        let result = run_squash_merge(repo, "feature/story-nope", "nope")
            .unwrap();

        assert!(!result.success, "merge of nonexistent branch should fail");
    }

    #[tokio::test]
    async fn merge_agent_work_conflict_does_not_break_master() {
        use std::fs;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        // Create a file on master.
        fs::write(repo.join("code.rs"), "fn main() {\n    println!(\"hello\");\n}\n").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "initial code"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Feature branch: modify the same line differently.
        Command::new("git")
            .args(["checkout", "-b", "feature/story-42_story_foo"])
            .current_dir(repo)
            .output()
            .unwrap();
        fs::write(repo.join("code.rs"), "fn main() {\n    println!(\"hello\");\n    feature_fn();\n}\n").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "feature: add fn call"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Master: add different line at same location.
        Command::new("git")
            .args(["checkout", "master"])
            .current_dir(repo)
            .output()
            .unwrap();
        fs::write(repo.join("code.rs"), "fn main() {\n    println!(\"hello\");\n    master_fn();\n}\n").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "master: add fn call"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Create story file in 4_merge.
        let merge_dir = repo.join(".story_kit/work/4_merge");
        fs::create_dir_all(&merge_dir).unwrap();
        fs::write(merge_dir.join("42_story_foo.md"), "---\nname: Test\n---\n").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "add story"])
            .current_dir(repo)
            .output()
            .unwrap();

        let pool = AgentPool::new(3001);
        let report = pool.merge_agent_work(repo, "42_story_foo").await.unwrap();

        // Master should NEVER have conflict markers, regardless of merge outcome.
        let master_code = fs::read_to_string(repo.join("code.rs")).unwrap();
        assert!(
            !master_code.contains("<<<<<<<"),
            "master must never contain conflict markers:\n{master_code}"
        );
        assert!(
            !master_code.contains(">>>>>>>"),
            "master must never contain conflict markers:\n{master_code}"
        );

        // The report should accurately reflect what happened.
        assert!(report.had_conflicts, "should report conflicts");
    }

    // ── process health monitoring tests ──────────────────────────────────────

    /// Demonstrates that the PTY read-loop inactivity timeout fires when no output
    /// is produced by the agent process within the configured window.
    ///
    /// A `HangingReader` simulates a hung agent process that never writes to the
    /// PTY master.  The test verifies that `recv_timeout` fires with a `Timeout`
    /// error — the signal that causes `run_agent_pty_blocking` to kill the child
    /// and return `Err("Agent inactivity timeout: …")`, which the error handler
    /// in `start_agent` converts into `AgentStatus::Failed`.
    #[test]
    fn pty_inactivity_timeout_kills_hung_agent() {
        struct HangingReader;
        impl std::io::Read for HangingReader {
            fn read(&mut self, _buf: &mut [u8]) -> std::io::Result<usize> {
                std::thread::sleep(std::time::Duration::from_secs(300));
                Ok(0)
            }
        }

        let (line_tx, line_rx) =
            std::sync::mpsc::channel::<std::io::Result<String>>();

        std::thread::spawn(move || {
            let buf_reader = BufReader::new(HangingReader);
            for line in buf_reader.lines() {
                if line_tx.send(line).is_err() {
                    break;
                }
            }
        });

        let timeout_dur = std::time::Duration::from_millis(100);
        let result = line_rx.recv_timeout(timeout_dur);

        assert!(
            matches!(
                result,
                Err(std::sync::mpsc::RecvTimeoutError::Timeout)
            ),
            "recv_timeout must fire when no PTY output arrives within the deadline"
        );
    }

    /// Demonstrates that the background watchdog detects Running agents whose
    /// backing tokio task has already finished (orphaned entries) and marks them
    /// as Failed, emitting an Error event so that `wait_for_agent` unblocks.
    #[tokio::test]
    async fn watchdog_detects_orphaned_running_agent() {
        let pool = AgentPool::new(3001);

        let handle = tokio::spawn(async {});
        tokio::time::sleep(std::time::Duration::from_millis(20)).await;
        assert!(handle.is_finished(), "task should be finished before injection");

        let tx =
            pool.inject_test_agent_with_handle("orphan_story", "coder", AgentStatus::Running, handle);
        let mut rx = tx.subscribe();

        pool.run_watchdog_once();

        {
            let agents = pool.agents.lock().unwrap();
            let key = composite_key("orphan_story", "coder");
            let agent = agents.get(&key).unwrap();
            assert_eq!(
                agent.status,
                AgentStatus::Failed,
                "watchdog must mark an orphaned Running agent as Failed"
            );
        }

        let event = rx.try_recv().expect("watchdog must emit an Error event");
        assert!(
            matches!(event, AgentEvent::Error { .. }),
            "expected AgentEvent::Error, got: {event:?}"
        );
    }

    // ── remove_agents_for_story tests ────────────────────────────────────────

    #[test]
    fn remove_agents_for_story_removes_all_entries() {
        let pool = AgentPool::new(3001);
        pool.inject_test_agent("story_a", "coder-1", AgentStatus::Completed);
        pool.inject_test_agent("story_a", "qa", AgentStatus::Failed);
        pool.inject_test_agent("story_b", "coder-1", AgentStatus::Running);

        let removed = pool.remove_agents_for_story("story_a");
        assert_eq!(removed, 2, "should remove both agents for story_a");

        let agents = pool.list_agents().unwrap();
        assert_eq!(agents.len(), 1, "only story_b agent should remain");
        assert_eq!(agents[0].story_id, "story_b");
    }

    #[test]
    fn remove_agents_for_story_returns_zero_when_no_match() {
        let pool = AgentPool::new(3001);
        pool.inject_test_agent("story_a", "coder-1", AgentStatus::Running);

        let removed = pool.remove_agents_for_story("nonexistent");
        assert_eq!(removed, 0);

        let agents = pool.list_agents().unwrap();
        assert_eq!(agents.len(), 1, "existing agents should not be affected");
    }

    // ── reap_expired_agents tests ────────────────────────────────────────────

    #[test]
    fn reap_expired_agents_removes_old_completed_entries() {
        let pool = AgentPool::new(3001);

        // Inject a completed agent with an artificial old completed_at.
        {
            let (tx, _) = broadcast::channel::<AgentEvent>(64);
            let key = composite_key("old_story", "coder-1");
            let mut agents = pool.agents.lock().unwrap();
            agents.insert(
                key,
                StoryAgent {
                    agent_name: "coder-1".to_string(),
                    status: AgentStatus::Completed,
                    worktree_info: None,
                    session_id: None,
                    tx,
                    task_handle: None,
                    event_log: Arc::new(Mutex::new(Vec::new())),
                    completion: None,
                    project_root: None,
                    log_session_id: None,
                    // Set completed_at 2 hours ago.
                    completed_at: Some(Instant::now() - std::time::Duration::from_secs(7200)),
                },
            );
        }
        // Inject a recently completed agent.
        pool.inject_test_agent("new_story", "coder-1", AgentStatus::Completed);
        // Inject a running agent (should not be reaped).
        pool.inject_test_agent("active_story", "coder-2", AgentStatus::Running);

        // Reap with a 1-hour TTL — only the old entry should be removed.
        let reaped = pool.reap_expired_agents(std::time::Duration::from_secs(3600));
        assert_eq!(reaped, 1, "should reap only the old completed entry");

        let agents = pool.list_agents().unwrap();
        assert_eq!(agents.len(), 2, "new_story and active_story should remain");
        assert!(
            agents.iter().all(|a| a.story_id != "old_story"),
            "old_story should have been reaped"
        );
    }

    #[test]
    fn reap_expired_agents_removes_old_failed_entries() {
        let pool = AgentPool::new(3001);

        // Inject a failed agent with an old completed_at.
        {
            let (tx, _) = broadcast::channel::<AgentEvent>(64);
            let key = composite_key("failed_old", "coder-1");
            let mut agents = pool.agents.lock().unwrap();
            agents.insert(
                key,
                StoryAgent {
                    agent_name: "coder-1".to_string(),
                    status: AgentStatus::Failed,
                    worktree_info: None,
                    session_id: None,
                    tx,
                    task_handle: None,
                    event_log: Arc::new(Mutex::new(Vec::new())),
                    completion: None,
                    project_root: None,
                    log_session_id: None,
                    completed_at: Some(Instant::now() - std::time::Duration::from_secs(7200)),
                },
            );
        }

        let reaped = pool.reap_expired_agents(std::time::Duration::from_secs(3600));
        assert_eq!(reaped, 1);
        assert!(pool.list_agents().unwrap().is_empty());
    }

    #[test]
    fn reap_expired_agents_skips_running_entries() {
        let pool = AgentPool::new(3001);
        pool.inject_test_agent("running_story", "coder-1", AgentStatus::Running);

        let reaped = pool.reap_expired_agents(std::time::Duration::from_secs(0));
        assert_eq!(reaped, 0, "running agents should never be reaped");
    }

    // ── archive + cleanup integration test ───────────────────────────────────

    #[tokio::test]
    async fn archiving_story_removes_agent_entries_from_pool() {
        use std::fs;

        let tmp = tempfile::tempdir().unwrap();
        let root = tmp.path();

        // Set up story in 2_current/
        let current = root.join(".story_kit/work/2_current");
        fs::create_dir_all(&current).unwrap();
        fs::write(current.join("60_story_cleanup.md"), "test").unwrap();

        let pool = AgentPool::new(3001);
        pool.inject_test_agent("60_story_cleanup", "coder-1", AgentStatus::Completed);
        pool.inject_test_agent("60_story_cleanup", "qa", AgentStatus::Completed);
        pool.inject_test_agent("61_story_other", "coder-1", AgentStatus::Running);

        // Verify all 3 agents exist.
        assert_eq!(pool.list_agents().unwrap().len(), 3);

        // Archive the story.
        move_story_to_archived(root, "60_story_cleanup").unwrap();
        pool.remove_agents_for_story("60_story_cleanup");

        // Agent entries for the archived story should be gone.
        let remaining = pool.list_agents().unwrap();
        assert_eq!(remaining.len(), 1, "only the other story's agent should remain");
        assert_eq!(remaining[0].story_id, "61_story_other");

        // Story file should be in 5_done/
        assert!(root.join(".story_kit/work/5_done/60_story_cleanup.md").exists());
    }

    // ── bug 154: merge worktree installs frontend deps ────────────────────

    /// When the feature branch has a `frontend/` directory, `run_squash_merge`
    /// must run `pnpm install` in the merge worktree before quality gates.
    /// This test creates a repo with a `frontend/package.json` and verifies that
    /// the merge output mentions the pnpm install step.
    #[cfg(unix)]
    #[test]
    fn squash_merge_runs_pnpm_install_when_frontend_exists() {
        use std::fs;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        // Add a frontend/ directory with a minimal package.json on master.
        let frontend = repo.join("frontend");
        fs::create_dir_all(&frontend).unwrap();
        fs::write(
            frontend.join("package.json"),
            r#"{"name":"test","version":"0.0.0","private":true}"#,
        )
        .unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "add frontend dir"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Create feature branch with a change.
        Command::new("git")
            .args(["checkout", "-b", "feature/story-154_test"])
            .current_dir(repo)
            .output()
            .unwrap();
        fs::write(repo.join("feature.txt"), "change").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "feature work"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Switch back to master.
        Command::new("git")
            .args(["checkout", "master"])
            .current_dir(repo)
            .output()
            .unwrap();

        let result =
            run_squash_merge(repo, "feature/story-154_test", "154_test").unwrap();

        // The output must mention pnpm install, proving the new code path ran.
        assert!(
            result.output.contains("pnpm install"),
            "merge output must mention pnpm install when frontend/ exists, got:\n{}",
            result.output
        );
    }

    /// When `pnpm install` fails in the merge worktree (e.g. no lockfile),
    /// the merge must abort cleanly — success=false, workspace cleaned up.
    #[cfg(unix)]
    #[test]
    fn squash_merge_aborts_when_pnpm_install_fails() {
        use std::fs;
        use tempfile::tempdir;

        let tmp = tempdir().unwrap();
        let repo = tmp.path();
        init_git_repo(repo);

        // Add a frontend/ directory with an invalid package.json that will
        // cause pnpm install --frozen-lockfile to fail (no lockfile present).
        let frontend = repo.join("frontend");
        fs::create_dir_all(&frontend).unwrap();
        fs::write(
            frontend.join("package.json"),
            r#"{"name":"test","version":"0.0.0","dependencies":{"nonexistent-pkg-xyz":"99.99.99"}}"#,
        )
        .unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "add frontend with bad deps"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Feature branch with a change.
        Command::new("git")
            .args(["checkout", "-b", "feature/story-154_fail"])
            .current_dir(repo)
            .output()
            .unwrap();
        fs::write(repo.join("change.txt"), "feature").unwrap();
        Command::new("git")
            .args(["add", "."])
            .current_dir(repo)
            .output()
            .unwrap();
        Command::new("git")
            .args(["commit", "-m", "feature"])
            .current_dir(repo)
            .output()
            .unwrap();

        // Switch back to master, record HEAD.
        Command::new("git")
            .args(["checkout", "master"])
            .current_dir(repo)
            .output()
            .unwrap();
        let head_before = String::from_utf8(
            Command::new("git")
                .args(["rev-parse", "HEAD"])
                .current_dir(repo)
                .output()
                .unwrap()
                .stdout,
        )
        .unwrap()
        .trim()
        .to_string();

        let result =
            run_squash_merge(repo, "feature/story-154_fail", "154_fail").unwrap();

        // pnpm install --frozen-lockfile should fail (no lockfile), merge aborted.
        assert!(
            !result.success,
            "merge should fail when pnpm install fails"
        );
        assert!(
            result.output.contains("pnpm install"),
            "output should mention pnpm install"
        );

        // Master HEAD must not have moved.
        let head_after = String::from_utf8(
            Command::new("git")
                .args(["rev-parse", "HEAD"])
                .current_dir(repo)
                .output()
                .unwrap()
                .stdout,
        )
        .unwrap()
        .trim()
        .to_string();
        assert_eq!(
            head_before, head_after,
            "master HEAD must not advance when pnpm install fails (bug 154)"
        );

        // Merge workspace should be cleaned up.
        assert!(
            !repo.join(".story_kit/merge_workspace").exists(),
            "merge workspace should be cleaned up after failure"
        );
    }

    // ── check_orphaned_agents return value tests (bug 161) ──────────────────

    #[tokio::test]
    async fn check_orphaned_agents_returns_count_of_orphaned_agents() {
        let pool = AgentPool::new(3001);

        // Spawn two tasks that finish immediately.
        let h1 = tokio::spawn(async {});
        let h2 = tokio::spawn(async {});
        tokio::time::sleep(std::time::Duration::from_millis(20)).await;
        assert!(h1.is_finished());
        assert!(h2.is_finished());

        pool.inject_test_agent_with_handle("story_a", "coder", AgentStatus::Running, h1);
        pool.inject_test_agent_with_handle("story_b", "coder", AgentStatus::Running, h2);

        let found = check_orphaned_agents(&pool.agents);
        assert_eq!(found, 2, "should detect both orphaned agents");
    }

    #[test]
    fn check_orphaned_agents_returns_zero_when_no_orphans() {
        let pool = AgentPool::new(3001);
        // Inject agents in terminal states — not orphaned.
        pool.inject_test_agent("story_a", "coder", AgentStatus::Completed);
        pool.inject_test_agent("story_b", "qa", AgentStatus::Failed);

        let found = check_orphaned_agents(&pool.agents);
        assert_eq!(found, 0, "no orphans should be detected for terminal agents");
    }

    #[tokio::test]
    async fn watchdog_orphan_detection_returns_nonzero_enabling_auto_assign() {
        // This test verifies the contract that `check_orphaned_agents` returns
        // a non-zero count when orphans exist, which the watchdog uses to
        // decide whether to trigger auto-assign (bug 161).
        let pool = AgentPool::new(3001);

        let handle = tokio::spawn(async {});
        tokio::time::sleep(std::time::Duration::from_millis(20)).await;

        pool.inject_test_agent_with_handle(
            "orphan_story",
            "coder",
            AgentStatus::Running,
            handle,
        );

        // Before watchdog: agent is Running.
        {
            let agents = pool.agents.lock().unwrap();
            let key = composite_key("orphan_story", "coder");
            assert_eq!(agents.get(&key).unwrap().status, AgentStatus::Running);
        }

        // Run watchdog pass — should return 1 (orphan found).
        let found = check_orphaned_agents(&pool.agents);
        assert_eq!(
            found, 1,
            "watchdog must return 1 for a single orphaned agent"
        );

        // After watchdog: agent is Failed.
        {
            let agents = pool.agents.lock().unwrap();
            let key = composite_key("orphan_story", "coder");
            assert_eq!(
                agents.get(&key).unwrap().status,
                AgentStatus::Failed,
                "orphaned agent must be marked Failed"
            );
        }
    }

    // ── kill_all_children tests ────────────────────────────────────

    /// Returns true if a process with the given PID is currently running.
    fn process_is_running(pid: u32) -> bool {
        std::process::Command::new("ps")
            .arg("-p")
            .arg(pid.to_string())
            .stdout(std::process::Stdio::null())
            .stderr(std::process::Stdio::null())
            .status()
            .map(|s| s.success())
            .unwrap_or(false)
    }

    #[test]
    fn kill_all_children_is_safe_on_empty_pool() {
        let pool = AgentPool::new(3001);
        // Should not panic or deadlock on an empty registry.
        pool.kill_all_children();
        assert_eq!(pool.child_killer_count(), 0);
    }

    #[test]
    fn kill_all_children_kills_real_process() {
        // GIVEN: a real PTY child process (sleep 100) with its killer registered.
        let pool = AgentPool::new(3001);

        let pty_system = native_pty_system();
        let pair = pty_system
            .openpty(PtySize {
                rows: 24,
                cols: 80,
                pixel_width: 0,
                pixel_height: 0,
            })
            .expect("failed to open pty");

        let mut cmd = CommandBuilder::new("sleep");
        cmd.arg("100");
        let mut child = pair
            .slave
            .spawn_command(cmd)
            .expect("failed to spawn sleep");
        let pid = child.process_id().expect("no pid");

        pool.inject_child_killer("story:agent", child.clone_killer());

        // Verify the process is alive before we kill it.
        assert!(
            process_is_running(pid),
            "process {pid} should be running before kill_all_children"
        );

        // WHEN: kill_all_children() is called.
        pool.kill_all_children();

        // Collect the exit status (prevents zombie; also ensures signal was sent).
        let _ = child.wait();

        // THEN: the process should be dead.
        assert!(
            !process_is_running(pid),
            "process {pid} should have been killed by kill_all_children"
        );
    }

    #[test]
    fn kill_all_children_clears_registry() {
        // GIVEN: a pool with one registered killer.
        let pool = AgentPool::new(3001);

        let pty_system = native_pty_system();
        let pair = pty_system
            .openpty(PtySize {
                rows: 24,
                cols: 80,
                pixel_width: 0,
                pixel_height: 0,
            })
            .expect("failed to open pty");

        let mut cmd = CommandBuilder::new("sleep");
        cmd.arg("1");
        let mut child = pair
            .slave
            .spawn_command(cmd)
            .expect("failed to spawn sleep");

        pool.inject_child_killer("story:agent", child.clone_killer());
        assert_eq!(pool.child_killer_count(), 1);

        // WHEN: kill_all_children() is called.
        pool.kill_all_children();
        let _ = child.wait();

        // THEN: the registry is empty.
        assert_eq!(
            pool.child_killer_count(),
            0,
            "child_killers should be cleared after kill_all_children"
        );
    }
}