bestdeejay-design/agent-skills / skills/docs-system
docs-system
VerifiedMeta-guide + router for documentation: the two branches (docs-product for 'why/what', docs-project for 'how'), the phase order, completeness levels L1/L2/L3, and the completeness checklist. Load a branch directly when the task is clearly product or engineering; load this skill for the big picture, ordering and completeness gates. Triggers: 'документация', 'набор документации', 'каталог документов', 'документация структура', 'полная документация', 'docs catalog', 'documentation structure', 'какую документацию писать'.
| File | Kind | Size |
|---|---|---|
| ROADMAP.md | markdown | 4.7 KB |
| SKILL.md | skill md | 3.4 KB |
| evals/evals.json | config | 570 B |
| examples/example-monorepo/README.md | markdown | 5.8 KB |
| readme.ru.md | markdown | 1.1 KB |
| references/completeness.md | markdown | 3.1 KB |
| references/levels.md | markdown | 5.0 KB |
| references/order.md | markdown | 3.4 KB |
| skill.json | config | 1.7 KB |
references/order.md
View on GitHub# Fill Order — from idea to complete documentation set
The order in which documents are created matters. It is the difference between a
folder of files and a coherent, drivable system. Follow these phases **in order**.
Later phases often reveal gaps in earlier ones — that's normal and healthy.
> **RU:** порядок создания файлов важен: сначала продуктовая документация, потом
> контракты, карта — последней. Поздние фазы вскрывают дыры в ранних — это норма.
## Golden rules
1. **Product before engineering** — VISION/PRD come before ARCHITECTURE. You
cannot design «how» before you know «why» and «what».
2. **Machine truth before code** — contracts (OpenAPI/AsyncAPI) are written
**before** implementation, not reverse-engineered from it.
3. **The map comes last** — `docs/REFERENCE.md` is written last, because it
describes everything above it. Writing it last forces a full verification.
## Phases
**Phase 1 — Product foundation**
- `docs/VISION.md` — 1 paragraph: what it is, who it's for, problem solved,
out-of-scope. If you can't write this, **stop**: clarify the idea first.
- `docs/PRD.md` (draft) — goals/non-goals, top stories, success metrics.
- `docs/BACKLOG.md` (L2+) — ideas, deferred features, priorities.
**Phase 2 — Plan & architecture (drafts)**
- `docs/ROADMAP.md` — milestones by value (not effort), each with a proof.
- `docs/ARCHITECTURE.md` (draft) — components, communication, data ownership.
- ADR process starts: `docs/ADR/ADR-001.md` (L2+) or `docs/DECISIONS.md` (L1).
**Phase 3 — Contracts FIRST (API-first / L3)**
- `contracts/openapi/*.yaml`, `contracts/asyncapi/events.yaml` **before any route
is written**. Prevents «the agent invented a route nobody planned».
**Phase 4 — Feature catalog**
- `docs/FEATURES.md` (L2+): ✅/📋 per feature/module.
- L3·microservices: scaffold services; types mirror the contracts.
**Phase 5 — Cross-module behavior (if >1 module)**
- `docs/SAGA.md`: trigger, steps, compensation, retries, DLQ.
**Phase 6 — Tests**
- `docs/TEST_CASES.md` (L2+) + contract-conformance tests.
- L1: a `TEST_PLAN.md` is allowed; grows to L2 TEST_CASES later.
**Phase 7 — Dev guide**
- `docs/DEV_GUIDE.md` (L2+): prerequisites, run, env, migrations, add-a-module.
- README «Quick start» must list real commands.
**Phase 8 — What running taught you**
- `docs/TROUBLESHOOTING.md` (L2+): errors E1…En + startup checklist.
- `docs/IMPROVEMENTS.md` (L2+): known issues, doc-vs-fact drift, fix plan.
- `AGENT.md`, `DELIVERY.md` (L2+): agent runbook + delivery gate.
**Phase 9 — Entry point + community (public repos)**
- `ENTRY.md` (L2+): navigation «topic → file» + onboarding checklist.
- `CONTRIBUTING.md`, `CODE_OF_CONDUCT.md`, `SECURITY.md`.
**Phase 10 — The map (LAST)**
- `docs/REFERENCE.md` — written **last**: card for every doc, hierarchy of truth,
drift table. Then run `references/completeness.md` until it passes.
## Rules that hold throughout
- **Update, don't recreate.** Each phase updates earlier docs.
- **One authoritative home per fact.** On conflict — the hierarchy of truth decides.
- **Touch a doc → touch its card.** Any edit implies `REFERENCE.md` maintenance in
the drift table.
- **Smallest adequate level.** L1 that can grow is better than a premature full
catalog. Growth rules live in `references/levels.md`.Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file). | Stars: 198k | Updated | Claude CodeCodex | Unverified | Install |
| doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |