Claude Skills by bgauryy

Use when planning, running, grading, or reporting the by-hand Octocode research benchmark — pairwise matchups (Octocode anchor vs one baseline: gh+RTK, gh+Headroom, or plain gh) over markdown questions, with a fresh isolated runner agent per (question, arm, pass), one blind judge per question grading two answers X/Y in randomized order, and an orchestrator that summarizes accuracy/quality/workflow/characters. Results measured in total characters through the model (model-in delivered + model-o...

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.

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 sensor. Not for ordinary ship checks where 'tests passed' is enough.

Use when posing, animating, or explaining a 3D human skeleton/manikin: anatomical joints, ROM clamps, walk/run/dance/backflip sequences, Three.js viewer, or agent-driven WebMCP figure control (reach/look_at). Phrases like pose the manikin, skeleton scheme, animate a walk cycle. Not for general 3D scenes, physics/ragdoll, IK foot-locking, or mocap import.

Use when extracting or mapping public web content into a local cited corpus: scrape or crawl a URL/docs site, pull tables/pricing/product fields, diagnose blocked or thin pages, or answer from saved pages. Phrases like scrape this URL, crawl the docs, build a corpus, extract pricing. Prefer keyless fetch; ask before hosted spend. Live clicks/HAR/perf → octocode-chrome-devtools.

Use when working on Agent Skills (SKILL.md folders): find, compare, rate, review, create, improve, repair, install, or sync skills from local workspaces, registries, or remote sources. For wording-only prompt polish without folder ops, use octocode-prompt-optimizer.

Load when synthesizing a new local skill from research. Why: plan + approve before writing. If fetching a remote skill first, load `references/fetch-remote.md`. Shape the lobby with `references/skill-anatomy.md` + `references/skill-improve.md` (workflows in `SKILL.md`, no overlaps).

Use when spawning workers or offloading work: Task/subagents, specialist handoffs, A2A peers, sealed packets, coordinate/synthesize, challenge techniques (rubber-duck, interview, mimic-flow, red-team, blind review, consensus), or local Ollama one-shots to save tokens. Measuring keep/discard KPIs → octocode-graph-eval.

Use when the user asks to design, review, implement, debug, or evaluate an agentic workflow or AI-agent harness: MCP tools/resources/prompts, multi-agent routing or handoffs, agent memory/context/cache, Zod/JSON-schema protocols, human gates, observability, evals, or production safety. Do not use for ordinary app code, generic prompt writing, or model comparison unless an agentic flow boundary is involved.

Idea brainstorming and validation grounded in evidence. Triggers on "brainstorm", "is this worth building", "has anyone built X", "validate my idea", "check if X exists", "research this idea", "what are the prior-art options for Y". Researches GitHub, npm/PyPI, and the web in parallel, then synthesizes a decision-ready brief — not code or designs.

Use Chrome DevTools Protocol (CDP) for browser debugging, inspection, and automation when DevTools-grade evidence is needed: network, console, performance, DOM/CSS, screenshots/PDF, security, storage, auth-gated, live-page, and source-traced findings. Opens or attaches to Chrome, runs sandboxed CDP scripts, and keeps sessions reusable. Prefer lighter browser tools for simply opening a page.

Use `octocode-cli` subcommands to execute Octocode MCP tools from a terminal without wiring MCP. Use when the user asks to "run octocode from shell", "use octocode without MCP", "call githubSearchCode from CLI", or wants a one-off GitHub code/file/PR search in the terminal.

High-level design-system and UI architecture generator for existing or new projects. Uses Octocode MCP local tools first, then creates a dynamic (not rigid) DESIGN.md covering visual language, styling strategy, component architecture, framework constraints, accessibility, performance, responsive behavior, and implementation guidance.

This skill should be used when the user asks to "generate documentation", "document this project", "create docs", "write documentation", "update documentation", "document all APIs", "generate onboarding docs", "create developer docs", or needs comprehensive codebase documentation. Orchestrates parallel AI agents to analyze code and produce documentation files.

System-engineering skill for codebase understanding, bug investigation, refactors, PR safety, architecture review, and RFC validation. Enforces Clean Architecture and Clean Code with AST, LSP, and scanner evidence. Produces a flows / boundaries / architecture-health artifact with file:line citations before recommending action.

Interactive step-by-step installer for Octocode tools on macOS and Windows. Use when the user asks to "install octocode", "set up octocode", "configure octocode mcp", "get started with octocode", "install octocode-cli", "octocode setup", or needs help with GitHub auth, IDE MCP config, or skills installation.

Researches what is new in AI, developer tools, web platform, security, and notable repositories. Use when the user asks for whats-new, latest updates, recent releases, tech news, AI news, changelogs, repo updates, or trend scanning.

Use when the user asks to "plan & implement", "plan this work", "research & build", "plan auth/API/work", or needs a multi-step pipeline from understanding through implementation. Flow is Understand → Research → Plan → Implement → Verify. For design documents or technical proposals without implementation, use octocode-rfc-generator instead.

This skill should be used when the user asks to "optimize this prompt", "improve this SKILL.md", "make this prompt more reliable", "fix my agent instructions", "review this AGENTS.md", "strengthen this prompt", "my agent keeps skipping steps", "add enforcement to instructions", or needs to transform weak prompts into reliable, enforceable agent protocols. Uses a 6-step gated flow (with Fast/Full modes), command strengthening, gate injection, and failure mode analysis.

This skill should be used when the user asks to "review a PR", "review pull request", "PR review", "check this PR", "analyze PR changes", "review PR #123", "what''s wrong with this PR", "is this PR safe to merge", "review my changes", "review local changes", "review my code", "review staged changes", "review my diff", or needs expert code review with architectural analysis, defect detection, and security scanning. Supports both remote PRs and local changes (staged/unstaged). Uses Octocode MCP...

Use when the user asks to "research code", "how does X work", "where is Y defined", "who calls Z", "trace code flow", "find usages", "explore this library", "understand the codebase", or needs deep code exploration with HTTP-based tool orchestration. For direct MCP tool research without the HTTP server, use octocode-researcher instead.

Primary research skill — use when the user asks to research, search, explore, find, trace, investigate, or understand code. Triggers include "find X", "where is Y defined?", "explore this dir", "trace definitions", "find usages", "how does X work?", "who calls Z?", "search for X", "research this library", "find PRs", "what package does X?", "understand this flow", "investigate this bug", "what changed?", or any code exploration/discovery need — local or external. Uses Octocode MCP tools direc...

Research-driven RFC and design document generator. Use when the user asks to "create an RFC", "write a design doc", "propose a migration", "how should we architect X", "evaluate options for X", "write a technical proposal", "compare approaches", or needs a technical decision document before coding. Outputs a validated RFC with research evidence, alternatives, recommendation, and implementation plan. For planning with implementation, use octocode-plan instead.

This skill should be used when the user asks to "roast my code", "review code brutally", "find code sins", "what's wrong with my code", "shame my code", "critique this code", "find antipatterns", "code quality roast", or wants entertaining but actionable code criticism with severity-ranked fixes. Delivers brutally honest roasts with file:line citations and redemption paths.

Use this skill when the user asks to find, evaluate, preview, install, rate, review, score, improve, refactor, or synthesize Agent Skills (the `SKILL.md` folder format) across GitHub, local skill folders, and skill marketplaces. Covers searching for a skill for a task, deep-diving a candidate, installing one or more skills into one or more agents at user or project scope, rating or reviewing an existing SKILL.md, refactoring a skill, or creating a new local skill from researched patterns. Do ...

Generates polished multi-file HTML presentations. Six-phase flow: brief → research → outline → design → implementation → review. Each slide is a standalone HTML file loaded via iframe. Use when asked to 'create slides', 'make a presentation', 'generate HTML slides', 'build a deck', or turn notes/docs/code into a polished presentation.

Render an Octocode MCP usage dashboard from `${OCTOCODE_HOME}/stats.json` or `~/.octocode/stats.json`. Use when the user asks to show Octocode stats, usage, tokens/chars saved, cache hits, errors, rate limits, or visualize `stats.json`.