octocode-benchmark
Use when planning, running, grading, or reporting the by-hand Octocode research benchmark — pairwise matchups (Octocode anchor vs one baseline: gh+RTK,…
Use when docs are missing, wrong, stale, or badly written, or need a copyedit against the Google style guide: README, API reference, runbook, CONTRIBUTING, changelog, onboarding, AGENTS.md/CLAUDE.md, ADR, migration guide, Diátaxis or whole-codebase restructure, docstrings, alt
$ npx -y skills add bgauryy/octocode-mcp --skill octocode-documentation --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/octocode-documentationContext preview
The summary Claude sees to decide when to auto-load this skill.
Use when docs are missing, wrong, stale, or badly written, or need a copyedit against the Google style guide: README, API reference, runbook, CONTRIBUTING, changelog, onboarding, AGENTS.md/CLAUDE.md, ADR, migration guide, Diátaxis or whole-codebase restructure, docstrings, alt
name: octocode-documentation description: "Use when docs are missing, wrong, stale, or badly written, or need a copyedit against the Google style guide: README, API reference, runbook, CONTRIBUTING, changelog, onboarding, AGENTS.md/CLAUDE.md, ADR, migration guide, Diátaxis or whole-codebase restructure, docstrings, alt text, prose linting. Not for code, commits, or marketing copy. Code investigation → octocode-research; SKILL.md folders → octocode-skills."
Evidence-backed docs for humans and agents, written to the Google developer documentation style guide. Classify first. Gate writes. Prefer durable cross-refs over code dumps.
`UNDERSTAND → RESEARCH → CLASSIFY → OUTLINE GATE → WRITE → STYLE → VERIFY`
UNDERSTAND names the deliverable, audience, and target paths. Compress when the request already names targets and type. Expand when claims need verification. A copyedit request starts at STYLE. Answer a single-term question ("is `allows you to` okay?") straight from `assets/google-word-list.tsv` — quote the guidance and stop.
Stop when: outline gate awaits answer; write+style+verify finishes; a word-list lookup answered the question; a missing fact makes the doc dishonest to write (otherwise mark "Not verified in repository" and continue); conventions conflict; user cancels.
| Mode | Deliverable | Route order | |---|---|---| | agent-docs | `AGENTS.md`, nested agent instructions, `CLAUDE.md` symlink | `modes.md` → `evidence-research.md` → `agents-md.md` → `agent-readable.md` → `write-verify.md` → `style-lint.mjs` | | human-docs | README, tutorial, how-to, reference, explanation, runbook | `modes.md` → `evidence-research.md` → `diataxis.md` → `agent-readable.md` → `write-verify.md` → `style-lint.mjs` | | adr | Architecture decision record | `modes.md` → `evidence-research.md` → `adr.md` → `write-verify.md` → `style-lint.mjs` | | codebase-pack | Multi-file docs set | `modes.md` → plan the file set → gate once → per file: `diataxis.md` → `write-verify.md` → `style-lint.mjs` | | style-pass | Edited text, or a style review report | `style-index.md` → the owning `style-*.md` → `style-lint.mjs` → `style-review.md` for a report someone else acts on |
| Phase | When | Read | |---|---|---| | CLASSIFY | Choosing mode or audience | `references/modes.md` | | RESEARCH | Gathering or verifying repository facts | `references/evidence-research.md` | | CLASSIFY | Choosing the Diátaxis type for human-docs, or reviewing one | `references/diataxis.md` | | WRITE | Writing or updating agent instruction files | `references/agents-md.md` | | WRITE | Recording a decision | `references/adr.md` | | WRITE | Cross-refs, density, durability — read before the first line | `references/agent-readable.md` | | OUTLINE GATE, WRITE, VERIFY | Outline gate, write steps, style pass, verify checklist | `references/write-verify.md` | | STYLE | Any wording, formatting, or terminology question — maps every guide topic to its owner | `references/style-index.md` |
| Ask | Read | |---|---| | Which reference owns this topic — every guide topic, one row each | `references/style-index.md` | | Prose: tone, person, voice, tense, grammar, one specific word, abbreviations, jargon, translation, inclusive terms | `references/style-voice.md`, `references/style-grammar.md`, `references/style-words.md`, `references/style-abbreviations.md`, `references/style-global.md`, `references/style-inclusive.md` | | Page shape: headings, lists, numbered steps, notices, tables, footnotes, figures, alt text | `references/style-structure.md`, `references/style-procedures.md`, `references/style-blocks.md`, `references/style-images.md` | | Mechanics: bold/italic/code choice, capitalization, filenames, markup, punctuation, numbers, dates, units | `references/style-format.md`, `references/style-punctuation.md`, `references/style-numbers.md` | | Technical text: code font, samples, command syntax, placeholders, example values, UI wording, link text | `references/style-code.md`, `references/style-cli.md`, `references/style-examples.md`, `references/style-ui.md`, `references/style-links.md` | | Claims and reference text: time words, superlatives, product names, trademarks, third-party text, docstrings | `references/style-claims.md`, `references/style-api.md` | | Producing a review someone else acts on, or checking a
Evidence-first code research for AI agents and developers. Octocode researches your local code and external code alike (GitHub repositories, PRs, npm) with one toolset: ripgrep + AST search, trees, precise reads, and LSP.
Repo: bgauryy/octocode-mcp
Use when planning, running, grading, or reporting the by-hand Octocode research benchmark — pairwise matchups (Octocode anchor vs one baseline: gh+RTK,…
Use when an idea needs disciplined exploration before building: generate options, test whether it is worth building, map adjacent solutions, challenge…
Use when a live page needs Chrome DevTools/CDP evidence: network failures, console errors, performance, DOM/CSS actionability, screenshots/PDF,…
Use when you need a measurable keep/discard loop — goal→KPI, baseline vs target, held-out checks, eval suites, or don't-stop-till-done against a runnable…
Use when posing, animating, or explaining a 3D human skeleton/manikin: anatomical joints, ROM clamps, walk/run/dance/backflip sequences, Three.js viewer, or…
Use when an agent prompt, tool schema, policy, or handoff needs to get clearer, safer, easier to trigger, cheaper in context, or measurable against real…