菜单
Use when writing or updating docs: README, API docs, runbooks, AGENTS.md, CLAUDE.md, ADRs, Diátaxis restructuring, or codebase documentation for humans or coding agents.
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
| name | octocode-documentation |
| description | Use when writing or updating docs: README, API docs, runbooks, AGENTS.md, CLAUDE.md, ADRs, Diátaxis restructuring, or codebase documentation for humans or coding agents. |
Evidence-backed docs for humans and agents. Classify first. Gate writes. Prefer durable cross-refs over code dumps.
UNDERSTAND → RESEARCH → CLASSIFY → OUTLINE GATE → WRITE → VERIFY
Compress when targets and type are named. Expand when claims need verification.
AGENTS.md as a docs index (links + non-obvious rules), not a content dump.AGENTS.md, or paste large code blocks — link the source.Stop when: outline gate awaits answer; write+verify finishes; fact missing (report gap); user cancels.
Load only what the step needs:
references/modes.md when choosing mode or audience.references/evidence-research.md when gathering or verifying repo facts.references/diataxis.md when writing or reviewing human-docs.references/agents-md.md when writing or updating agent instruction files.references/adr.md when recording a decision.references/agent-readable.md before WRITE (cross-refs, density, durability).references/write-verify.md for outline gate, write steps, and verify checklist.references/references.md during normal tasks (creation audit only).octocode-research.SKILL.md → octocode-skills.octocode-documentation-writer if installed.Use when coordinating work in a shared repo — multi-agent or solo across sessions: collision avoidance, handoffs, verification debt, durable memory/wiki, hooks setup/debug, and repo learning before planning, editing, reviewing, or testing.
Use when an idea needs disciplined exploration before building: generate options, test worth-building, map adjacent solutions, challenge assumptions, narrow scope, or choose Build RFC / Prototype / Narrow / Park — even without saying brainstorm.
Use when browser debugging needs Chrome DevTools evidence: network, console, performance, DOM/CSS, screenshots/PDF, security, cookies/storage, or auth-gated live pages via CDP — not just opening a URL.
Use when measuring whether a change helped: ACCEPT/REVERT, keep/discard, goal→KPI contracts, eval suites, graders, held-out checks, benchmarks, or TDD failing-case-first.
Use when saving tokens with a local Ollama worker while the cloud agent keeps tools and judgment: summarize, extract, classify, translate, article digest after fetch, draft, checklist, vision caption, map-reduce, or phrases like ollama worker, local model, offload, use ollama.
Use when an agent prompt, skill, tool schema, policy, or handoff needs to get clearer, safer, easier to trigger, cheaper in context, or measurable against real behavior.