القائمة
Use when implementing GitHub issues for skill or plugin development through TDD-governed phases with iterative planning, implementation, and review loops. Use this whenever the issue involves SKILL.md files, agent prompts, hook scripts, or plugin workflows.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
| name | develop |
| description | Use when implementing GitHub issues for skill or plugin development through TDD-governed phases with iterative planning, implementation, and review loops. Use this whenever the issue involves SKILL.md files, agent prompts, hook scripts, or plugin workflows. |
| allowed-tools | ["Task","Skill","Read","Write","Edit","Bash","Grep","Glob","TodoWrite","WebFetch","AskUserQuestion"] |
| user-invocable | true |
Implement skill development issues through iterative planning, TDD-governed implementation, and review cycles.
/plugin-dev instead/plugin-dev:create instead| Phase | Action | Exit Condition |
|---|---|---|
| 0. Setup | MUST create feature branch | On feature branch (NOT main/master) |
| 1. Context | Read issue + detect skill files | Requirements clear, skill files noted |
| 2. Plan | Draft implementation plan | Plan complete |
| 3-4. Validate | Gemini review loop (max 5) | APPROVED status |
| 4.5. Baseline (RED) | TODO: #163 — capture baseline behavior | Baseline captured (or skipped pending #163) |
| 5-7. Implement (GREEN) | Write minimal skill addressing failures | All tasks done + criteria met + tests passing + validation passing |
| 7.5. Verify (REFACTOR) | TODO: #163 — compare against baseline | Benchmark passes (or skipped pending #163) or escalated |
| 8-9. Review | Council review (max 3) | APPROVED status |
| 10. Finalize | Commit + PR | PR created |
| 11. Cleanup | Clean temps, delete branch | Branch deleted after merge |
| Change Size | Criteria | Consultants |
|---|---|---|
| Trivial | <10 lines, no logic changes | Skip review → Phase 10 |
| Small | 1-2 files, simple logic | council:gemini-consultant only |
| Medium | 3-5 files, moderate complexity | council:gemini-consultant + council:codex-consultant |
| Large | 6+ files, architectural impact | /council skill |
During Phase 1, detect and note these skill-related files for domain-specific handling:
SKILL.md — Skill activation definitions (YAML frontmatter + instructions)references/*.md — Workflow reference files (e.g., references/WORKFLOW.md)agents/*.md — Agent/subagent definitionshooks/hooks.json — Hook config (declares hook types, matchers, and script references)scripts/ — Hook scripts (executables referenced by hooks.json via ${CLAUDE_PLUGIN_ROOT}/scripts/...)Use this skill when the user says: "develop skill", "skill dev", "plugin develop", "develop issue", "skill workflow", "implement skill issue", "TDD skill development".
Read references/WORKFLOW.md now — it contains the detailed step-by-step procedures for every phase, including TDD extension points, loop limits, and escalation rules. Follow it exactly.
Consult external AI council (Gemini 3.5 Flash, Codex, Qwen, GLM-5.2, Kimi) for thorough reviews and consensus-driven decisions. Use ONLY when explicitly invoked with "/council" or when user says "consult the council", "invoke council", or "council review". Do NOT auto-trigger on generic phrases like "thorough review".
PR babysitter: monitors CI status, auto-rebases when behind, auto-fixes CI where possible, delegates review comment handling to dlc:pr-check, and re-requests review after fixes. Designed for /loop usage with Remote Control.
Council review reference data — expertise weights, structured response format, scoring thresholds, and false positive taxonomy. Background knowledge for council subagents.
Audit review checklist for a repo: cluster recurring review-comment themes across merged PRs in a window, diff against docs/code-review-checklist.md, and open a PR proposing new entries with PR traceability. Designed for /schedule monthly cadence. Pass --dry-run to print clusters without opening a PR. Trigger phrases: audit review checklist, update review checklist, cluster PR findings.
Multi-agent development workflow using Agent Teams. Supports five modes: plan (architect teammate + PM teammate debate → plan.md), dev (developer teammate + code-tester teammate + qa-tester teammate + reviewer teammate iterate → code), full (plan → approval gate → dev), auto (plan → dev, no gate), and bugfix (tester + developer + reviewer TDD triad → fix + PR). Use when tasks benefit from collaborative agent teammates with peer messaging.
You MUST use this skill before /pm when work is ambiguous, scope is unclear, or multiple approaches exist. Explores problem space, proposes 2-3 approaches with trade-offs, and produces an approved spec document that feeds into both issue creation (/pm) and implementation planning (/cdt). Triggers: brainstorm, explore idea, think through, design approach, what should we build, explore options, weigh approaches, compare solutions, needs help deciding between alternatives.