メニュー
Implement one Solo todo — write code, verify it, commit with the commit skill, save a result note, and close the todo. Use when acting as a worker subagent executing a single well-scoped Solo todo.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Find deepening opportunities in a codebase, informed by CONTEXT.md and docs/adr/. Records findings in Solo scratchpads and can turn selected refactors into Solo todos. Use when improving architecture, finding refactoring opportunities, or making a codebase more testable and AI-navigable.
Solo-native planning workflow. Use when asked to "plan", "brainstorm", "design", "create a plan", "I want to build", or "let's build". Covers both orchestrating a plan (scout → planner → workers → review) and acting as the interactive planner subagent.
Review code changes for quality, security, and correctness. Use when asked to "review", "code review", "check my changes", or when acting as a reviewer subagent. Produces a structured review with prioritized findings, saved to a Solo scratchpad when one is provided.
Fast codebase reconnaissance — map existing code, conventions, and patterns for a task. Use when asked to "scout", "explore the codebase for", or when acting as a scout subagent gathering context for a planner or worker.
Write clear Solo todos that workers can execute without losing architectural intent. Use when asked to "create todos", "write todos", "break into tasks", "plan todos", or create work items from a plan. Ensures each todo has outcomes, examples, constraints, and verification criteria.
Read this skill before making git commits
| name | worker |
| description | Implement one Solo todo — write code, verify it, commit with the commit skill, save a result note, and close the todo. Use when acting as a worker subagent executing a single well-scoped Solo todo. |
Execute exactly one well-scoped Solo todo: implement, verify, commit if code changed, save your result, mark the todo complete, and stop.
Do not redesign, re-plan, expand scope, or spawn subagents.
Your task references a Solo todo id. Read it through solo_tool because todo_get is gateway-only:
solo_tool({
action: "call",
name: "todo_get",
arguments: { todo_id: <id>, include_comments: true }
})
If a plan scratchpad or scout scratchpad is referenced, read it with scratchpad_read.
Before editing, verify the todo includes:
If required context is missing, do not guess. Use todo_update to append a clear ## Worker Blocker note to the todo body, save a short blocked note to your scratchpad, leave the todo incomplete, and stop.
Workers are normally run sequentially by the parent, so do not use todo locks.
Read files before editing. Keep changes minimal and focused. Follow existing project conventions.
Run the smallest meaningful verification: targeted tests, typecheck, build, smoke command, or endpoint/page check as appropriate. Capture command output for your result note.
Never claim success without verification evidence. If verification cannot run, explain why.
If code changed, read and follow ~/.pi/agent/skills/commit/SKILL.md before committing. Make one polished commit for this todo unless the task explicitly says not to commit.
Save this to the Solo scratchpad if one was provided, otherwise include it in your final message:
# Worker Result: Todo <id>
## Summary
[What changed]
## Files Changed
- `path` — [why]
## Verification
```bash
<command>
[relevant output]
<sha> — subject[Any follow-up or blocker]
### 7. Complete
If the todo is done, call `todo_complete({ todo_id: <id>, completed: true })`. Your final message should mention the todo id, scratchpad name/id, commit sha if any, and verification command.