菜单
Use when executing implementation plans with independent tasks in the current session. Dispatches a fresh subagent per task with two-stage review (spec compliance, then code quality).
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Systematic architecture review workflow. Use before major feature work on an unfamiliar or inherited codebase, before scaling milestones, or as a periodic health check. Maps the system as-built (not as-documented), dispatches parallel specialist reviews, and produces an evidence-backed findings report with severity verdicts.
The meta-skill. Loaded at every session start. Defines which A Team skills and agents MUST be used for which situations. If a skill or agent applies, you do not have a choice — you must use it.
Conversational onboarding for A Team. Detects ROADMAP.md, extracts context, generates INIT.md without requiring technical knowledge. Invoked automatically by the orchestrator when INIT.md is missing.
Root cause analysis using the Five Whys technique. Use when a bug persists despite surface fixes, a failure recurs, or a process keeps breaking. Ensures the fix targets the root, not the symptom.
Hard gate before implementing any API endpoint or service interface. Requires a written, reviewed contract (OpenAPI 3.x, protobuf, or GraphQL schema) before any implementation code is written. Prevents breaking changes, misaligned clients, and undocumented behaviour.
Use BEFORE any creative work — creating features, building components, adding functionality. Explores user intent, requirements, and design before implementation. Hard-gated: no code until user approves the spec.
| name | subagent-driven-development |
| description | Use when executing implementation plans with independent tasks in the current session. Dispatches a fresh subagent per task with two-stage review (spec compliance, then code quality). |
Execute a plan by dispatching fresh subagents per task, with two-stage review after each.
Core principle: Fresh subagent per task + two-stage review (spec then quality) = high quality, fast iteration.
Continuous execution: Do not pause to check in between tasks. Execute all tasks without stopping. The only reasons to stop: BLOCKED status you cannot resolve, genuine ambiguity, or all tasks complete.
finishing-a-development-branch skill| Status | Action |
|---|---|
| DONE | Proceed to spec compliance review |
| DONE_WITH_CONCERNS | Read concerns, address if correctness/scope issues |
| NEEDS_CONTEXT | Provide missing context, re-dispatch |
| BLOCKED | Assess: context problem → more context; reasoning → higher model; too large → break up; wrong plan → escalate |
Select the lowest tier that can handle the task:
| Scope | Claude Code | OpenAI | |
|---|---|---|---|
| 1-2 files, clear spec | claude-haiku-4-5 | gpt-4.1-mini | gemini-2.0-flash-lite |
| Multi-file, integration concerns | claude-sonnet-4-6 | gpt-4.1 | gemini-2.5-flash |
| Design judgment, broad codebase | claude-opus-4-8 | o3 / gpt-5.5 | gemini-2.5-pro |
Spec compliance MUST be ✅ before starting code quality review. Never skip either review stage.
Uses: writing-plans (creates the plan), finishing-a-development-branch (completes work)
Subagents use: test-driven-development, code-review patterns