Agent skill
quanzhankaifa
Use when the user wants a doc-first, end-to-end backend/full-stack project workflow with phase documents, confirmations, and traceability (inputs may be DOC/PDF/MD).
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/quanzhankaifa
SKILL.md
Quanzhankaifa (Full-Stack Project, Doc-First)
Overview
Use a doc-first delivery flow: from requirements to release, write every step into Markdown, pause for user confirmation after each doc, and prove alignment with the docs before any implementation work.
Skill Check (Required, Superpowers-Inspired)
- Before any action or clarification, check if other skills apply
- If any apply, invoke them first; process skills take priority
- Common triggers:
- Brainstorming: new features, unclear requirements, or design choices
- Systematic debugging: bugs, test failures, unexpected behavior
- Test-driven development: any code changes or new behavior
- Verification before completion: claiming work is done or fixed
Pre-Phase: Intent Clarification (Brainstorming Mode)
- Ask one question at a time; prefer multiple-choice when possible
- Clarify purpose, constraints, success criteria, non-goals, stakeholders
- When options exist, propose 2-3 approaches with trade-offs, recommend one
- Record Q/A, options, and decisions in
00-discovery.md - Update
00-index.mdand stop for confirmation after00-discovery.md
Core Workflow (Doc-First)
Produce or update a file at each phase:
0. Discovery and options -> 00-discovery.md
- Requirements ->
01-requirements.md - Scope and success ->
02-scope.md - Architecture ->
03-architecture.md - Modules and domain ->
04-module-design.md - Database ->
05-db-design.md - API ->
06-api-spec.md - Implementation plan ->
07-implementation-plan.md - Test plan ->
08-test-plan.md - Release and rollback ->
09-release-plan.md - Decisions and risks ->
10-decision-log.md,11-risk-log.md - Traceability ->
12-traceability.md
Working Directory + Logging Rules
- Default docs folder:
docs/project/ - Single entry index:
docs/project/00-index.md - Update the index every step (phase, done, open questions, next step)
- End each doc with: Summary + Open questions + Next step
- All doc content must be written in Chinese; keep filenames as specified
- After each doc is written, stop and ask the user to confirm or correct it
- If any Open questions remain, do not proceed until the user answers or defers them explicitly
- Read order each session:
00-index.md-> current phase doc -> related docs (minimal load)
Confirmation Gate (Required)
- After producing a phase doc, ask for confirmation in chat and wait
- Show a short "Doc alignment" list that references the doc filename and key bullets you followed
- Copy Open questions into the message and ask the user to answer or confirm deferral
- If the doc is not in Chinese, fix it before asking for confirmation
Traceability (Doc-to-Work Evidence)
- Maintain
12-traceability.mdwith a simple table:- Columns: Requirement | Design/Module | DB/API | Implementation/Tests | Status
- Update this table after each phase and before any coding
- Before coding, cite the relevant doc sections and the traceability rows you are implementing
Language Rule
- Use Chinese templates in
assets/and keep all headings, bullets, and narrative in Chinese across00-12docs - Comments should be written in Chinese unless the existing codebase clearly uses another language
Step-by-Step Execution
0) Discovery and options
- Use the Pre-Phase clarification loop and write
00-discovery.md - Stop and confirm
1) Requirements intake
- If input is DOCX/PDF: extract key points first, then write to
01-requirements.md - Output: features, non-functional requirements, constraints, acceptance criteria
- Stop and confirm
2) Architecture and modules
- Output: system boundary, core services/modules, data flow, dependencies
- Write to
03-architecture.mdand04-module-design.md - Stop and confirm
3) Database and API design
- Output: ER relations, tables, indexes, enums, constraints
- Output: API list, request/response, auth, error codes
- Write to
05-db-design.mdand06-api-spec.md - Stop and confirm
4) Implementation plan
- Output: milestones, task breakdown, dependencies, risks
- Write to
07-implementation-plan.md - Stop and confirm
5) Testing and release
- Testing: unit, integration, acceptance, performance
- Release: rollout, rollback, monitoring
- Write to
08-test-plan.mdand09-release-plan.md - Stop and confirm
Context Minimization
- Start by reading
00-index.md - Load only docs related to the current phase
- Close the loop by updating docs and reporting the changes
Quick Reference
| Phase | Output | Key questions |
|---|---|---|
| Discovery | 00-discovery.md | What are goals, constraints, options? |
| Requirements | 01-requirements.md | What is the goal and boundary? |
| Scope | 02-scope.md | What is MVP and what is out? |
| Architecture | 03-architecture.md | How does data and control flow? |
| Modules | 04-module-design.md | Clear responsibilities and deps? |
| Database | 05-db-design.md | Tables, indexes, constraints? |
| API | 06-api-spec.md | Request/response/auth/errors? |
| Plan | 07-implementation-plan.md | Milestones, dependencies, risks? |
| Testing | 08-test-plan.md | Coverage and acceptance? |
| Release | 09-release-plan.md | Rollback and monitoring? |
| Traceability | 12-traceability.md | Does code map to docs? |
Common Mistakes
- Keep decisions only in chat, not in docs
- Mix requirements and design in one file
- Write code first and backfill docs later
- Skip confirmation after a doc is written
- Code without showing doc alignment or traceability
- Write docs in English or mixed language
- Leave complex logic without comments
- Skip the discovery phase or ask multiple questions at once
Resources
- Doc templates:
references/api_reference.md - MD template:
assets/project-template.md - DOC template:
assets/project-template.doc - Alibaba Java spec (summary):
references/alibaba-java-spec.md - Code comment template:
references/code-comment-template.md
Handoff Rules
- For implementation, create
07-implementation-plan.mdbefore coding - For code changes, follow test-driven development
- For unexpected behavior, run systematic debugging first
- When writing code, add thorough comments for all non-trivial logic, control flow, and edge cases
- Use
references/code-comment-template.mdfor file/module/class/function/logic-block comments - Before and after coding, cite the docs you followed and update
12-traceability.md
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?