task-planner
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseTask Planner
任务规划器(Task Planner)
This skill helps you turn ambiguous or large requests into a clear, sequenced plan with ownership and verification.
本Skill可帮助你将模糊或庞大的需求转化为清晰、有序且明确责任人与验证方式的规划方案。
Outputs To Produce
需生成的输出内容
- A short problem statement
- Assumptions and open questions
- A step-by-step plan with measurable outcomes
- Risks and rollback/containment options
- Test and verification steps
- 简短的问题陈述
- 假设前提与待解决问题
- 具备可衡量成果的分步计划
- 风险项与回退/应对方案
- 测试与验证步骤
Decomposition Technique
任务分解技巧
Split work into thin vertical slices:
- One slice should be mergeable on its own
- Each slice should include tests or validation
- Prefer smallest unit that reduces uncertainty
Common slice types:
- Spec tests (lock requirements, CLI flags, edge cases)
- Implementation changes (small, isolated)
- Documentation updates
- Observability and traceability improvements
将工作拆分为细粒度的垂直切片:
- 每个切片需可独立合并
- 每个切片应包含测试或验证环节
- 优先选择能降低不确定性的最小单元
常见切片类型:
- 规格测试(锁定需求、CLI参数、边缘场景)
- 实现变更(小型、独立的修改)
- 文档更新
- 可观测性与可追溯性优化
Dependency And Priority Rules
依赖与优先级规则
- Identify blockers first (missing API, failing tests, permissions).
- Order by "unblocks others" and "reduces uncertainty".
- For multi-agent work, keep interfaces stable and define contracts (inputs/outputs) per slice.
- 首先识别阻塞项(缺失的API、失败的测试、权限问题等)。
- 按照“解除其他任务阻塞”和“降低不确定性”的顺序排列任务。
- 多Agent协作时,保持接口稳定,并为每个切片定义契约(输入/输出)。
Multi-Agent Assignment
多Agent任务分配
When delegating:
- Specify the exact deliverable (files, tests, output format)
- Specify constraints (no commits, branch constraints, time limits)
- Specify the acceptance test (what must pass)
Example delegation message:
Please write tests for <feature>.
Constraints:
- Do not change implementation yet
- Use pytest
Acceptance:
- Tests fail before implementation
- Tests cover edge cases: <...>分配任务时:
- 明确指定具体交付物(文件、测试、输出格式)
- 明确约束条件(禁止提交代码、分支限制、时间限制)
- 明确验收测试标准(需通过的验证项)
示例分配消息:
Please write tests for <feature>.
Constraints:
- Do not change implementation yet
- Use pytest
Acceptance:
- Tests fail before implementation
- Tests cover edge cases: <...>TodoList Integration
TodoList集成
Use a TodoList to track execution:
- Keep items small, phrased as outcomes
- Include "verify" tasks (tests, manual checks)
- Mark completed immediately to prevent drift
Example TodoList:
- Add failing tests for <behavior>
- Implement <behavior> behind tests
- Run targeted tests
- Update docs
- Run full test suite
使用TodoList跟踪执行进度:
- 任务项需保持粒度细小,以成果为表述方式
- 包含“验证”类任务(测试、人工检查)
- 完成后立即标记,防止进度偏差
示例TodoList:
- 为<behavior>添加失败测试用例
- 在测试驱动下实现<behavior>
- 运行针对性测试
- 更新文档
- 运行完整测试套件
Progress Reporting
进度汇报
For status updates, report:
- What is done (with tests run)
- What is next
- What is blocked and by what
状态更新需包含:
- 已完成事项(含已执行的测试)
- 下一步计划
- 阻塞项及阻塞原因