task-processor
Process a task list one sub-task at a time with pause/confirm gates, test/commit protocol, and file tracking.
- License 许可
- MIT
- 最近更新
- 2026-09-02
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- task-processor
- category
- workflow
- description
- Process a task list one sub-task at a time with pause/confirm gates, test/commit protocol, and file tracking.
02 资源文件 2
- 参考文档reference.md257 B
- 参考文档skill-bundle-manifest.json1003 B
03 SKILL.md 正文
Task Processor
Task Implementation
- One sub-task at a time. Do not start the next sub-task until user says "yes" or "y".
- Completion protocol:
- When finishing a sub-task, mark it
[x]. - If all sub-tasks under a parent are
[x], then:- Run full test suite.
- If tests pass: stage changes (
git add .). - Clean up temporary files/code.
- Commit using conventional commit style with multi-
-mmessages describing changes and referencing task/PRD.git commit -m "feat: add payment validation logic" -m "- Validates card type and expiry" -m "- Adds unit tests for edge cases" -m "Related to T123 in PRD"
- After commit, mark the parent task
[x].
- When finishing a sub-task, mark it
- Stop after each sub-task and wait for user go-ahead.
Task List Maintenance
- Update the task list as you work: mark
[x], add tasks if needed. - Maintain the "Relevant Files" section with one-line purpose per file.
Production Completion Criteria
A sub-task is NOT complete until:
- Implementation works: The code functions correctly
- Tests pass: All related tests pass
- End-to-end works: The change works in the full user flow
- No blocking issues: Any discovered issues are fixed
Critical Rule: Fix Issues In-Place
When you encounter an issue while implementing or testing:
- DO NOT mark the sub-task complete with caveats
- DO NOT say "this works, but there's an unrelated issue"
- DO NOT defer issues to future tasks
- DO fix the issue as part of the current work
- DO re-test after each fix
- DO only mark complete when it works end-to-end
Before requesting user approval to proceed:
- Verify the sub-task works in isolation
- Verify it works in the broader context
- If you found and fixed additional issues, mention them
- Confirm the feature is production-ready
AI Instructions
- Regularly update the task list after finishing significant work.
- Follow completion protocol strictly.
- Add newly discovered tasks.
- Keep "Relevant Files" accurate and up to date.
- Before starting, check which sub-task is next.
- After implementing a sub-task, update the file and pause for approval.
- Validate end-to-end before marking any sub-task complete.
- Fix all blocking issues in-place, never defer them.
References
- See
reference.md. - See
.claude/agents/production-validator.mdfor validation guidance.
04 同仓库更多 Skills
fixture-skill
Secret-free recovery fixture skill.
user-story-generator
Generate standalone user stories for features without creating a full PRD. Perfect for backlog grooming and story refinement.
user-journey-mapper
Map user flows and journeys through your application for UX design and workflow visualization, independent of testing.
transcript-query
暂无简介
test-plan-generator
Generate comprehensive E2E test plans from PRDs, task lists, or descriptions with environment-aware prerequisites and issue tracking.
tasklist-generator
Generate high-level tasks and gated sub-tasks from a PRD, with agent assignments, commit/PR strategy, and relevant files.