Skills Library

Create product/project showcase videos using Remotion (React). Takes project description + screenshots → generates compositions → renders MP4. Use when asked to make demo videos, product showcases, or animated project walkthroughs.

Extract knowledge from YouTube videos into BrainLayer. Use when user shares a YouTube link or asks to process/watch/extract from a video. Chains exa (transcript) -> brain_digest (entities/relations) -> brain_store (conclusions). Works with any YouTube URL.

Use when debriefing meetings, practicing presentations, QA testing with voice, or capturing insights to Obsidian. Covers voice drilling, coaching, capture. NOT for: simple TTS announcements (use voice_speak directly).

Build and polish presentations using Michal's workshop method + Oren Efraim's rules. Guides users through premise, framing, drafting, polishing, opening/closing, and practice. Works with voice (VoiceLayer) or text.

Content creation and publishing for ClaudeGolem across platforms (Soltome, blog, social)

The complete PR loop — branch, implement, test, commit, push, PR, WAIT FOR REVIEW, fix, merge, cleanup. Includes PR creation and review comment fetching. Use whenever creating a PR or finishing work. This is NOT optional. Every change goes through this loop. No exceptions.

Use when preparing a PR for QA review. Generates manual testing checklist from git diff. Covers test plan, QA checklist, testing before merge. NOT for: automated tests (write those separately), code reviews (use coderabbit).

Use when ready to commit changes. Runs CodeRabbit review first, then commits if review passes. Supports Ralph mode for atomic commit + criterion marking. Covers commit, ralph commit, atomic commit. NOT for: pushing or creating PRs (use pr-loop).

Use when needing multi-agent verification of complex work. Runs parallel critique agents until consensus. Covers verification, consensus, multi-agent review, validate work. NOT for: simple code reviews (use coderabbit), single-reviewer tasks.

Use when reviewing uncommitted changes, preparing PRs, checking for security issues, or verifying code quality. Runs AI code reviews via CLI. Covers code review, PR review, security scan, secrets scan. NOT for: runtime debugging (use debugger), test execution (run tests directly).

Use when needing semantic code navigation - find definitions, references, or callers. Covers LSP, go to definition, find references, hover, code intelligence. NOT for: text pattern searches (use grep), file discovery (use glob).

Use when starting isolated feature work. Creates git worktrees to prevent branch cross-contamination. Covers worktree, isolated development, parallel features, branch isolation. NOT for: simple branch switching (use git checkout), Linear-only operations (use linear).

Life admin assistant for health/habits, recruiting/jobs, freelancing/contracts, Israeli law, outreach/networking, and scheduling. Use when discussing daily planning, schedule creation, habit tracking, WHOOP data, job hunting, freelance contracts, Israeli business law, client management, or outreach emails. Also triggers for any conversation that references past coaching sessions or personal context that needs memory recall. Even seemingly simple requests ("build me a schedule", "check my WHOOP") benefit from this skill because coachClaude's value comes from persistent memory and accumulated context about the user's life, habits, and goals.

Use as fallback browser automation when Claude-in-Chrome MCP is unavailable. Covers browser control, navigation, screenshots, clicking, typing. NOT for: headless testing (use Playwright). Claude Code users should prefer MCP first.

Iterative Figma-to-implementation pixel-perfect verification loop. Use when implementing or refining UI from Figma designs. Drills on screenshots, comparing Figma vs implementation, fixing one thing at a time until 3 consecutive checks pass. Covers figma iteration, pixel perfect, design verification, ui drilling, figma comparison. NOT for: fetching Figma specs only (use figma-workflow docs), creating new components from scratch without a reference design.

Run external CLI agents (Gemini, Cursor, Codex, Kiro, Claude) via cmux panes. Workers split in current workspace. Audits/research open in a new named workspace. All agents are interactive and visible.

7 interview modes for technical interview preparation with Elo tracking

---

Spawn AI agents in cmux panes — Claude workers as splits, audits/research as surfaces. Covers Claude, Cursor, Gemini, Codex, Kiro. Includes monitoring, prompt delivery, and collab patterns. Use this skill whenever the user mentions cmux agents, terminal agents, split agents, multi-agent orchestration, or wants to spawn AI workers in visible terminal panes.

Set up the golems ecosystem for the first time on a new machine. Checks CLI dependencies, wires MCP servers, creates skill symlinks. Use when: "set up golems", "install golems", "new machine setup", "wire skills". NOT for daily usage.

Use when managing secrets, credentials, API keys, or vault operations. Supports Environments (Beta) for .env mounting. Covers 1password, secrets, op, vault, migrate, credentials. NOT for: non-secret config (use regular config files).

Use when working with Convex backend - dev server, deployment, function execution, data import/export. Wraps npx convex CLI. Covers convex, backend, serverless, functions, schema. NOT for: other databases (Supabase, Firebase, etc.), frontend-only work.

Use when doing git operations, creating PRs, or managing GitHub issues. Provides gh CLI commands. Covers git, github, commits, PRs, issues, branches, worktrees. NOT for: Linear issues (use linear), code reviews (use coderabbit).

Use when running inside cmux terminal to control panes, splits, browser, sidebar, and coordinate multi-agent workflows. Covers split panes, notifications, browser automation, agent-to-agent messaging via cmux send. NOT for: regular terminal operations (use Bash), non-cmux sessions.

Deploy and manage the golems cloud-worker on Railway. Use when deploying backend changes, checking logs, managing env vars, or restarting services. Wraps `railway` CLI. Covers railway, deploy, cloud-worker, redeploy, logs, variables. NOT for: Vercel deployments, frontend, Supabase.

First-time onboarding wizard for the golems ecosystem. Detects installed AI CLIs, configures workspace, sets up skills. Use when setting up golems for the first time, reconfiguring, or onboarding a new machine.

Collect stats from golems repo, update etanheyman.com content, detect dead references, and auto-PR. Use when stats drift (nightly, post-merge, manual). NOT for editorial rewrites.

Ecosystem-wide status collection and orientation. Use when returning to work, starting a new session, when the user says "where were we", "what's the status", "catch me up", "what happened", or any time you need to understand the current state across projects. This is NOT the repo-scoped /catchup (git diffs) — this is the orchestrator-level "what's happening across the whole ecosystem." Also use when you need to prepare a research prompt — this skill gathers all the context needed to write a detailed, self-contained prompt for a research agent.

Send Telegram notification to user when task completes

Use when reaching a "Commit:" criterion in Ralph stories. Atomically commits and marks criterion checked. Covers ralph commit, atomic commit, commit criterion. NOT for: regular git commits (use git directly), commits outside Ralph workflow.

Audit and fix per-project AI context hygiene. Compares loaded skills, MCPs, hooks, and agents against the project's context profile in ~/.golems/config.yaml. Reports wasted tokens and generates .claude/settings.local.json + CLAUDE.md containerization. Use when setting up a new project, debugging context bloat, or running maintenance sweeps.

Check email status, run manual triage, view recent scores. Use when asking about emails or wanting to check urgent messages.

LinkedIn writing coach — find topics from code history, draft posts following 2026 algorithm rules, review drafts. NOT auto-posting.

Run ecosystem health checks — MCP connections, BrainLayer stats, skill evals, friction scans. Use this skill when asked about ecosystem health, maintenance checks, skill monitoring, 'is everything working', 'run a health check', 'what's broken', or when proactively auditing the system. Also triggers for 'maintenance Claude', 'ecosystem audit', 'skill eval', 'MCP status', or 'BrainLayer health'. Run this before and after major changes to catch regressions.

Use when requesting or receiving code review. Covers dispatching reviewers, reading feedback, classifying issues, pushing back on wrong suggestions, and implementing fixes. Supersedes superpowers:requesting-code-review and superpowers:receiving-code-review.

Multi-agent Figma screen decomposition and component build pipeline. Use when implementing multiple screens from Figma designs, breaking down screens into components, mapping to existing component library, and coordinating parallel builds. Triggers on "decompose screens", "build from Figma", "screen components", "figma swarm", "component mapping", "design to components", or when the user shares multiple Figma screen node IDs for implementation. Also use when re-checking for design drift against Figma.

Use when wanting to see what skills are available. Lists installed skills with descriptions. Covers list skills, search skills, discover skills, what skills. NOT for: using a specific skill (invoke that skill directly).

Scaffold and execute folder-based multi-phase plans with async agent collaboration. Use when planning large features, multi-PR workflows, or coordinating parallel agent work across phases. NOT for: single-file changes, simple bugs, quick tasks.

Use when creating new golem-powers skills, editing existing skills, or verifying skills work. Covers create skill, write skill, skill template, skill structure. NOT for: using existing skills (invoke them directly), superpowers skills (different structure).

MANDATORY before reporting on any file contents, test results, agent outputs, or audit findings. If you haven't Read() it, you don't know what's in it. Period. Use when summarizing results, reporting on agent work, or claiming anything is "green" or "complete."

Use when planning a sprint transition, archiving completed work, or resetting the PRD for a fresh start. Archives completed PRD stories to docs.local/. Covers archive, cleanup, sprint transition, reset PRD. NOT for: deleting stories mid-sprint (use prd-manager), viewing PRD status (use ralph-status).

Deep web research orchestrator. Routes research tasks to the best backend — internal subagents, CLI agents (Gemini/Cursor), or the researcher subagent. Use when asked to research, investigate, compare, find alternatives, or deep-dive into any topic. Covers web research, company research, code pattern research, and pre-implementation research.

Use when returning to work after any break — auto-detects depth. Short break (hours): reads only uncommitted changes. Long break (48h+) or context overflow: reads all branch changes vs main. Covers catchup, context recovery, refresh, rebuild understanding. NOT for: mid-task exploration (use Read/Grep directly).

Use when needing current API references, function signatures, or library usage patterns. Looks up documentation via Context7 API. Covers docs lookup, library documentation, API reference, how to use library. NOT for: web search (use WebSearch), project-specific code (read the codebase).

Use when auditing an UNFAMILIAR codebase for the first time — architecture mapping, undocumented features, configuration gaps. NOT for catching up on your own branch (use catchup). Always searches BrainLayer first before touching files.

Use when accessing Obsidian vault notes - reading, searching, listing, or organizing notes. Covers obsidian, notes, vault, ideas, diary, memos. NOT for: general file operations outside the vault.