planning-and-task-breakdown
Use when shaping a rough idea, specifying an ambiguous feature, stress-testing a plan against code and project decisions, or turning an approved design into executable vertical slices.
- License 许可
- MIT
- 最近更新
- 2026-08-20
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- planning-and-task-breakdown
- description
- Use when shaping a rough idea, specifying an ambiguous feature, stress-testing a plan against code and project decisions, or turning an approved design into executable vertical slices.
02 资源文件 2
- 参考文档references/ADR-FORMAT.md2.7 KB
- 参考文档references/CONTEXT-FORMAT.md2.2 KB
03 SKILL.md 正文
Planning and Task Breakdown
<HARD-GATE> For rough, ambiguous, or design-heavy work, do not write code or an implementation plan until the design is approved. If requirements and cause are already concrete and reversible, skip discovery and use incremental implementation. </HARD-GATE>
1. Discover the Design
Classify unknowns before proposing:
- Known knowns: stated facts.
- Known unknowns: ask the user.
- Unknown knowns: show 2–4 cheap variants or point at a reference.
- Unknown unknowns: surface decision criteria and highest-risk blindspots.
For live alternatives, show trade-offs and recommend one. Ask one question at a time about the load-bearing architecture, data, UX, security, cost, or scope decision. State safe assumptions and their later verification. Use YAGNI against speculative scope, never against an explicit requirement.
2. Write the Contract
Spec before code. The spec is the contract. Minimum spec:
## Goal
<observable outcome>
## Non-goals
<explicit exclusions>
## Acceptance criteria
- [ ] <checkable behavior>
## Open questions
- <must resolve before implementation>
Tests derive from acceptance criteria. Spec changes are explicit scope changes, recorded before implementation continues. Keep WHAT separate from the implementation plan's HOW.
3. Stress-Test Material Decisions
Inspect code, tests, CONTEXT.md / CONTEXT-MAP.md, and project ADRs under docs/adr/ first. If exploration can answer a question, explore instead of asking. Cross-check claims against code and call out conflicting glossary terms or vague or overloaded language.
A blocking question must be:
- Material: answer changes architecture, scope, UX, data, security, cost, or criteria.
- Grounded: cite the concrete uncertainty.
- Answerable: provide live options, a recommended answer, and default if the user does not care.
Probe 10x scale, cost of being wrong, failure modes, rollback, uncontrolled dependencies, forgotten non-goals, and simplest validation. Stop when answers are concrete or questions repeat; do not grill reversible trivia or attack the user.
Create glossary files lazily; keep CONTEXT.md domain-only, not an implementation spec. See references/CONTEXT-FORMAT.md. Offer a project ADR under docs/adr/ only when the decision is hard to reverse, surprising without context, and the result of a real trade-off; see references/ADR-FORMAT.md. Work-session decisions belong in <repo-root>/.pi/artifacts/DECISIONS.md under development-lifecycle.
4. Build the Executable Plan
Order independently verifiable vertical slices:
- Put the most-likely-to-change model/interface/UX or unknown integration first.
- Give each slice one observable outcome, verification command, risk, and clean rollback.
- Put mechanical rename/reformat/refactor work last.
- Name open questions that block a slice and stop conditions for parallel workers.
## Goal / Non-goals
## Slices
1. <vertical outcome> — verify: <command> — risk: <failure>
## Open questions
## Stop conditions
Reject plans that begin with horizontal setup, accept “looks right,” omit non-goals, or cannot verify slices independently.
04 同仓库更多 Skills
writing-skills
Use when creating new skills, editing existing skills, or verifying skills work before deployment - applies TDD to process documentation by testing with subagents before writing, iterating until bulletproof against rationalization. Includes complete pressure testing methodology.
verification-before-completion
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always
typescript-coding-standards
Use when writing, refactoring, or reviewing TypeScript code that needs strong domain modeling, typed errors, schema parsing, safe adapters, test seams, or maintainable module boundaries.
test-driven-development
Use when implementing a feature, bugfix, or behavior-changing refactor — require a failing behavior test before production code, then make the smallest change that passes.
source-driven-development
Use when unfamiliar libraries, dependency internals, external APIs, framework behavior, migrations, or current ecosystem guidance require official documentation, version-matched source, read the tests, and cited evidence.
security-and-hardening
Use when auditing security, implementing authentication or authorization, handling secrets, or validating data across external trust boundaries.