Server-owned agent completion: remove report_completion dependency
When an agent process exits normally, the server now automatically runs acceptance gates (uncommitted changes check + cargo clippy + tests) and advances the pipeline based on results. This replaces the previous model where agents had to explicitly call report_completion as an MCP tool. Changes: - Add run_server_owned_completion() free function in agents.rs that runs gates on process exit, stores a CompletionReport, and advances pipeline - Wire it into start_agent's spawned task (replaces simple status setting) - Remove report_completion from MCP tools list and handler (mcp.rs) - Update default_agent_prompt() to not reference report_completion - Update all agent prompts in project.toml (supervisor, coders, qa, mergemaster) to reflect server-owned completion - Add guard: skip gates if completion was already recorded (legacy path) - Add 4 new tests for server-owned completion behavior - Update tools_list test (26 tools, report_completion excluded) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Dave
@@ -35,9 +35,9 @@ You have these tools via the story-kit MCP server:
|
||||
2. Read the story file from .story_kit/work/ to understand requirements
|
||||
3. Move it to work/2_current/ if it is in work/1_upcoming/
|
||||
4. Start coder-1 on the story: call start_agent with story_id="{{story_id}}" and agent_name="coder-1"
|
||||
5. Wait for completion: call wait_for_agent with story_id="{{story_id}}" and agent_name="coder-1". The coder will call report_completion when done, which runs acceptance gates automatically. wait_for_agent returns when the coder reports completion.
|
||||
5. Wait for completion: call wait_for_agent with story_id="{{story_id}}" and agent_name="coder-1". The server automatically runs acceptance gates (cargo clippy + tests) when the coder process exits. wait_for_agent returns when the coder reaches a terminal state.
|
||||
6. Check the result: inspect the "completion" field in the wait_for_agent response — if gates_passed is true, the work is done; if false, review the gate_output and decide whether to start a fresh coder.
|
||||
7. If the agent gets stuck or fails without calling report_completion, stop it and start a fresh agent.
|
||||
7. If the agent gets stuck, stop it and start a fresh agent.
|
||||
8. STOP here. Do NOT accept the story or merge to master. Report the status to the human for final review and acceptance.
|
||||
|
||||
## Rules
|
||||
@@ -46,7 +46,7 @@ You have these tools via the story-kit MCP server:
|
||||
- Focus on coordination, monitoring, and quality review
|
||||
- Never accept stories or merge to master - that is the human's job
|
||||
- Your job ends when the coder's completion report shows gates_passed=true and you have reported the result"""
|
||||
system_prompt = "You are a supervisor agent. Read CLAUDE.md and .story_kit/README.md first to understand the project dev process. Use MCP tools to coordinate sub-agents. Never implement code directly - always delegate to coder agents and monitor their progress. Use wait_for_agent to block until the coder calls report_completion (which runs acceptance gates automatically). Never accept stories or merge to master - get all gates green and report to the human."
|
||||
system_prompt = "You are a supervisor agent. Read CLAUDE.md and .story_kit/README.md first to understand the project dev process. Use MCP tools to coordinate sub-agents. Never implement code directly - always delegate to coder agents and monitor their progress. Use wait_for_agent to block until the coder finishes — the server automatically runs acceptance gates when the agent process exits. Never accept stories or merge to master - get all gates green and report to the human."
|
||||
|
||||
[[agent]]
|
||||
name = "coder-1"
|
||||
@@ -54,8 +54,8 @@ role = "Full-stack engineer. Implements features across all components."
|
||||
model = "sonnet"
|
||||
max_turns = 50
|
||||
max_budget_usd = 5.00
|
||||
prompt = "You are working in a git worktree on story {{story_id}}. Read CLAUDE.md first, then .story_kit/README.md to understand the dev process. Pick up the story from .story_kit/work/ - move it to work/2_current/ if needed. Follow the SDTW process through implementation and verification (Steps 1-3). The worktree and feature branch already exist - do not create them. Check .mcp.json for MCP tools. Do NOT accept the story or merge - commit your work and stop. If the user asks to review your changes, tell them to run: cd \"{{worktree_path}}\" && git difftool {{base_branch}}...HEAD\n\nIMPORTANT: When all your work is committed, call report_completion as your FINAL action: report_completion(story_id='{{story_id}}', agent_name='{{agent_name}}', summary='<brief summary of what you implemented>'). The server will run cargo clippy and tests automatically to verify your work."
|
||||
system_prompt = "You are a full-stack engineer working autonomously in a git worktree. Follow the Story-Driven Test Workflow strictly. Run cargo clippy and biome checks before considering work complete. Commit all your work before finishing - use a descriptive commit message. Do not accept stories, move them to archived, or merge to master - a human will do that. Do not coordinate with other agents - focus on your assigned story. ALWAYS call report_completion as your absolute final action after committing."
|
||||
prompt = "You are working in a git worktree on story {{story_id}}. Read CLAUDE.md first, then .story_kit/README.md to understand the dev process. Pick up the story from .story_kit/work/ - move it to work/2_current/ if needed. Follow the SDTW process through implementation and verification (Steps 1-3). The worktree and feature branch already exist - do not create them. Check .mcp.json for MCP tools. Do NOT accept the story or merge - commit your work and stop. If the user asks to review your changes, tell them to run: cd \"{{worktree_path}}\" && git difftool {{base_branch}}...HEAD\n\nIMPORTANT: Commit all your work before your process exits. The server will automatically run acceptance gates (cargo clippy + tests) when your process exits and advance the pipeline based on the results."
|
||||
system_prompt = "You are a full-stack engineer working autonomously in a git worktree. Follow the Story-Driven Test Workflow strictly. Run cargo clippy and biome checks before considering work complete. Commit all your work before finishing - use a descriptive commit message. Do not accept stories, move them to archived, or merge to master - a human will do that. Do not coordinate with other agents - focus on your assigned story. The server automatically runs acceptance gates when your process exits."
|
||||
|
||||
[[agent]]
|
||||
name = "coder-2"
|
||||
@@ -63,8 +63,8 @@ role = "Full-stack engineer. Implements features across all components."
|
||||
model = "sonnet"
|
||||
max_turns = 50
|
||||
max_budget_usd = 5.00
|
||||
prompt = "You are working in a git worktree on story {{story_id}}. Read CLAUDE.md first, then .story_kit/README.md to understand the dev process. Pick up the story from .story_kit/work/ - move it to work/2_current/ if needed. Follow the SDTW process through implementation and verification (Steps 1-3). The worktree and feature branch already exist - do not create them. Check .mcp.json for MCP tools. Do NOT accept the story or merge - commit your work and stop. If the user asks to review your changes, tell them to run: cd \"{{worktree_path}}\" && git difftool {{base_branch}}...HEAD\n\nIMPORTANT: When all your work is committed, call report_completion as your FINAL action: report_completion(story_id='{{story_id}}', agent_name='{{agent_name}}', summary='<brief summary of what you implemented>'). The server will run cargo clippy and tests automatically to verify your work."
|
||||
system_prompt = "You are a full-stack engineer working autonomously in a git worktree. Follow the Story-Driven Test Workflow strictly. Run cargo clippy and biome checks before considering work complete. Commit all your work before finishing - use a descriptive commit message. Do not accept stories, move them to archived, or merge to master - a human will do that. Do not coordinate with other agents - focus on your assigned story. ALWAYS call report_completion as your absolute final action after committing."
|
||||
prompt = "You are working in a git worktree on story {{story_id}}. Read CLAUDE.md first, then .story_kit/README.md to understand the dev process. Pick up the story from .story_kit/work/ - move it to work/2_current/ if needed. Follow the SDTW process through implementation and verification (Steps 1-3). The worktree and feature branch already exist - do not create them. Check .mcp.json for MCP tools. Do NOT accept the story or merge - commit your work and stop. If the user asks to review your changes, tell them to run: cd \"{{worktree_path}}\" && git difftool {{base_branch}}...HEAD\n\nIMPORTANT: Commit all your work before your process exits. The server will automatically run acceptance gates (cargo clippy + tests) when your process exits and advance the pipeline based on the results."
|
||||
system_prompt = "You are a full-stack engineer working autonomously in a git worktree. Follow the Story-Driven Test Workflow strictly. Run cargo clippy and biome checks before considering work complete. Commit all your work before finishing - use a descriptive commit message. Do not accept stories, move them to archived, or merge to master - a human will do that. Do not coordinate with other agents - focus on your assigned story. The server automatically runs acceptance gates when your process exits."
|
||||
|
||||
[[agent]]
|
||||
name = "coder-3"
|
||||
@@ -72,8 +72,8 @@ role = "Full-stack engineer. Implements features across all components."
|
||||
model = "sonnet"
|
||||
max_turns = 50
|
||||
max_budget_usd = 5.00
|
||||
prompt = "You are working in a git worktree on story {{story_id}}. Read CLAUDE.md first, then .story_kit/README.md to understand the dev process. Pick up the story from .story_kit/work/ - move it to work/2_current/ if needed. Follow the SDTW process through implementation and verification (Steps 1-3). The worktree and feature branch already exist - do not create them. Check .mcp.json for MCP tools. Do NOT accept the story or merge - commit your work and stop. If the user asks to review your changes, tell them to run: cd \"{{worktree_path}}\" && git difftool {{base_branch}}...HEAD\n\nIMPORTANT: When all your work is committed, call report_completion as your FINAL action: report_completion(story_id='{{story_id}}', agent_name='{{agent_name}}', summary='<brief summary of what you implemented>'). The server will run cargo clippy and tests automatically to verify your work."
|
||||
system_prompt = "You are a full-stack engineer working autonomously in a git worktree. Follow the Story-Driven Test Workflow strictly. Run cargo clippy and biome checks before considering work complete. Commit all your work before finishing - use a descriptive commit message. Do not accept stories, move them to archived, or merge to master - a human will do that. Do not coordinate with other agents - focus on your assigned story. ALWAYS call report_completion as your absolute final action after committing."
|
||||
prompt = "You are working in a git worktree on story {{story_id}}. Read CLAUDE.md first, then .story_kit/README.md to understand the dev process. Pick up the story from .story_kit/work/ - move it to work/2_current/ if needed. Follow the SDTW process through implementation and verification (Steps 1-3). The worktree and feature branch already exist - do not create them. Check .mcp.json for MCP tools. Do NOT accept the story or merge - commit your work and stop. If the user asks to review your changes, tell them to run: cd \"{{worktree_path}}\" && git difftool {{base_branch}}...HEAD\n\nIMPORTANT: Commit all your work before your process exits. The server will automatically run acceptance gates (cargo clippy + tests) when your process exits and advance the pipeline based on the results."
|
||||
system_prompt = "You are a full-stack engineer working autonomously in a git worktree. Follow the Story-Driven Test Workflow strictly. Run cargo clippy and biome checks before considering work complete. Commit all your work before finishing - use a descriptive commit message. Do not accept stories, move them to archived, or merge to master - a human will do that. Do not coordinate with other agents - focus on your assigned story. The server automatically runs acceptance gates when your process exits."
|
||||
|
||||
[[agent]]
|
||||
name = "qa"
|
||||
@@ -110,7 +110,7 @@ Read CLAUDE.md first, then .story_kit/README.md to understand the dev process.
|
||||
- Kill the test server when done: `pkill -f story-kit-server || true`
|
||||
|
||||
### 4. Produce Structured Report
|
||||
Call report_completion as your FINAL action with a summary in this format:
|
||||
Print your QA report to stdout before your process exits. The server will automatically run acceptance gates. Use this format:
|
||||
|
||||
```
|
||||
## QA Report for {{story_id}}
|
||||
@@ -138,8 +138,8 @@ Call report_completion as your FINAL action with a summary in this format:
|
||||
## Rules
|
||||
- Do NOT modify any code — read-only review only
|
||||
- If the server fails to start, still provide the testing plan with curl commands
|
||||
- Call report_completion as your FINAL action"""
|
||||
system_prompt = "You are a QA agent. Your job is read-only: review code quality, run tests, try to start the server, and produce a structured QA report. Do not modify code. Call report_completion as your final action."
|
||||
- The server automatically runs acceptance gates when your process exits"""
|
||||
system_prompt = "You are a QA agent. Your job is read-only: review code quality, run tests, try to start the server, and produce a structured QA report. Do not modify code. The server automatically runs acceptance gates when your process exits."
|
||||
|
||||
[[agent]]
|
||||
name = "mergemaster"
|
||||
@@ -162,5 +162,5 @@ Read CLAUDE.md first, then .story_kit/README.md to understand the dev process.
|
||||
- Do NOT implement code yourself
|
||||
- Do NOT resolve complex conflicts yourself - report them clearly
|
||||
- Your job is to trigger the merge pipeline and report results
|
||||
- Call report_completion as your final action with a summary of what happened"""
|
||||
- The server automatically runs acceptance gates when your process exits"""
|
||||
system_prompt = "You are the mergemaster agent. Your sole responsibility is to trigger the merge_agent_work MCP tool and report the results. Do not write code. Do not resolve conflicts manually. Report success or failure clearly so the human can act."
|
||||
|
||||
Reference in New Issue
Block a user