Skill index

• chain-hub

  New Protected Since v0.3.0
  • automation
  • development

  Chain Hub monorepo and on-disk hub: hub path resolution, protected core vs user skills, registry buckets, IDE adapters, and chain init / setup / update / validate / status / list. Use for chain-hub development (cli, core, apps/web), explaining how assets sync to editors, hub migration, skills not appearing in an editor, mistakes with registry.yaml and core/ layout, or troubleshooting wrong binary, broken symlinks, and sync tools.

  chain add chain-hub

  Source

• chain-memory-rules

  New Protected Since v0.3.0
  • automation

  Chain Hub memory rules: chain memory init, chain memory sync, and chain capture --rule. Use for remembering preferences, workspace context, ~/chain-hub/memory, hub #brain (Memory tab; legacy #memory redirects), or syncing rules into Cursor, Claude Code, Windsurf, and other linked editors. Cross-editor where adapters support memory sync.

  chain add chain-memory-rules

  Source

• chain-setup

  New Protected Since v0.3.0
  • automation

  Chain Hub first-run and editor linking: chain init, chain setup, chain status, and first chain validate. Use for new installs, relinking after CLI upgrade, or when hub assets should appear in Cursor, Claude Code, Windsurf, and other linked editors. Cross-editor.

  chain add chain-setup

  Source

• chain-validate

  New Protected Since v0.3.0
  • automation
  • development

  Fix chain validate and hub validation errors: registry buckets, flat skills layout, ghost slugs, pack overlap, skill/workflow/agent/rule frontmatter. Use when validate fails or skills-registry.yaml drifts from disk.

  chain add chain-validate

  Source

• code-review

  New Protected Since v0.3.0
  • development
  • automation

  Five-dimension code review covering correctness, readability, security, performance, and tests. Use when reviewing pull requests, code changes, or when the user asks for a code review, says "review this", "check my code", or "bekijk mijn code". Produces structured feedback with severity labels so reviewers and authors can prioritize fixes. Cross-editor, works in Cursor, Claude Code, Windsurf, and any other AI editor.

  chain add code-review

  Source ★ 850

• commit

  Protected Since v0.3.0
  • automation
  • development

  Writes structured git commit messages following Conventional Commits. Use when the user asks to commit, says "commit this", mentions writing a commit message, or when git commit context is present. In the chain-hub monorepo, commits touching cli/ or apps/hub/ MUST use conventional types (feat, fix, revert, refactor, chore, docs, test). Only feat/fix/revert update cli/CHANGELOG.md via git-cliff. Produces type(scope): subject format with optional body, roadmap refs, and breaking change notation — cross-editor.

  chain add commit

  Source

• create-agent

  New Authoring Since v0.3.0
  • automation

  Create Chain Hub agents (~/chain-hub/agents/*.md) with name and description frontmatter. Use for subagents, delegation personas, or chain agents new. Cross-editor via chain setup.

  chain add create-agent

  Source

• create-hook

  Authoring Since v0.3.0
  • automation

  Create agent hooks for Cursor (hooks.json + scripts) or Claude Code (settings.json hooks blocks). Use when you want to run custom logic before or after agent events, gate or audit tool calls, or automate workflows triggered by agent actions.

  chain add create-hook

  Source

• create-rule

  Authoring Since v0.3.0
  • automation
  • web

  Create persistent AI rules: Cursor .mdc rules, Chain Hub rules/, or editor-specific instruction files (CLAUDE.md, AGENTS.md). Use when you want to add coding standards, project conventions, file-specific patterns, or asks about .cursor/rules/, rules/global.md, or AGENTS.md.

  chain add create-rule

  Source

• create-skill

  New Authoring Since v0.3.0
  • automation
  • web

  Guides users through creating effective agent skills (SKILL.md) for Cursor, Claude Code, Windsurf, Gemini/Antigravity, and other Chain Hub–linked editors. Use when you want to create, write, or author a new skill, or asks about skill structure, best practices, or SKILL.md format.

  chain add create-skill

  Source

• create-subagent

  Authoring Since v0.3.0
  • automation

  Deprecated — use create-agent for Chain Hub agents (~/chain-hub/agents/). Install with chain add create-agent. Kept for backward compatibility with older docs.

  chain add create-subagent

  Source

• create-workflow

  New Authoring Since v0.3.0
  • automation

  Create Chain Hub workflows (~/chain-hub/workflows/*.md) that pass chain validate. Use for procedures, checklists, chain workflows new, or hub #workflows. Cross-editor via chain setup.

  chain add create-workflow

  Source

• debug

  New Protected Since v0.3.0
  • development
  • automation

  Systematic debugging workflow for isolating and fixing root causes. Use when hitting an error, test failure, or unexpected behavior, or when the user says "why isn't this working", "this is broken", or "help me debug". Follows a four-step approach: reproduce → isolate → fix root cause → guard with a regression test. Cross-editor, works in Cursor, Claude Code, Windsurf, and any other AI editor.

  chain add debug

  Source

• editor-adapter-troubleshoot

  New Protected Since v0.3.0
  • automation
  • development

  Skills or rules missing in Cursor, Claude Code, Windsurf, or other linked editors. Use with chain status and chain setup; covers symlinks, adapters, core shadowing, and per-editor paths. Cross-editor.

  chain add editor-adapter-troubleshoot

  Source

• find-skills

  Protected Since v0.3.0
  • automation
  • web

  Discover and install agent skills for Chain Hub (registry + skills.sh). Use when the user asks how to do something ("how do I …"), wants to extend capabilities ("find a skill for …", "is there a skill that can …", "can you do …"), needs a workflow or domain pack, or mentions skill discovery, skills.sh, or npx skills.

  chain add find-skills

  Source ★ 1.2k

• hub-content-authoring

  New Authoring Since v0.3.0
  • automation

  Route Chain Hub authoring to the right skill: skills, rules, workflows, agents, or hooks in your hub. Use when the user wants to add hub content but has not specified the type. Ends with chain validate.

  chain add hub-content-authoring

  Source

• hub-learn

  New Protected Since v0.3.0
  • automation
  • development

  Chain Hub Learn pipeline: capture learnings, chain reflect, improve proposals, apply, validate, and rollback. Use for #brain (Pipeline tab), self-improvement, improving skills/workflows/rules/agents, hub-gardener, or continual learning. Never silently edit protected npm core.

  chain add hub-learn

  Source

• hub-sync

  New Protected Since v0.3.0
  • automation

  Sync your hub across machines with chain sync (Git). Use for chain sync init, chain sync status, multi-device hubs, or recovering from bad folder sync. Warns against Syncthing or Dropbox on the hub folder parallel to chain sync.

  chain add hub-sync

  Source

• migrate-to-skills

  Authoring Since v0.3.0
  • automation
  • web

  Convert Cursor 'applied intelligently' rules and slash commands to skill folders (SKILL.md). Optionally migrate Claude Code slash commands from .claude/commands. Use when migrating .mdc rules or .md commands into skills under Chain Hub or editor skill directories.

  chain add migrate-to-skills

  Source

• project-context-bootstrap

  New Protected Since v0.4.0
  • automation
  • development

  Bootstrap high-quality AI project context for greenfield or early-stage repos. Detects linked editors and scaffolds IDE-specific instruction files (CLAUDE.md, AGENTS.md, GEMINI.md, .cursor/rules/*.mdc, .windsurfrules, and other Chain Hub adapter targets). Pair with project-health for Claude Code permissions and context audits after bootstrap. Use when starting a new project, onboarding a repo to AI editors, or when the user asks for CLAUDE.md, AGENTS.md, or multi-editor project setup from day one. Cross-editor.

  chain add project-context-bootstrap

  Source

• registry-hygiene

  New Protected Since v0.3.0
  • automation

  Maintain skills-registry.yaml: buckets, github_sources, authorship.self, packs, deactivated skills, and pack/core overlap. Use when registering installs, cleaning ghost slugs, or fixing chain validate registry errors.

  chain add registry-hygiene

  Source

• skill-opportunity-mapper

  New Protected Since v0.4.0
  • automation
  • development

  Walks through a work process and extracts skill opportunities using four criteria (recurring, clear input/output, has rules, non-trivial). Produces a saved skill map and an impact-vs-effort build queue ready for create-skill. Use for yourself or 1:1 with a client or student. Triggers include "what skills should I build", "find skill opportunities", "map my process", "what could become a skill", "audit my workflow", "where can Claude help me", "I want to skillify my work", and "skill ideas from my process".

  chain add skill-opportunity-mapper

  Source

• skill-search-authoring

  New Protected Since v0.3.0
  • automation
  • web

  Write skill descriptions and tags for Chain Hub registry and marketplace search. Use when authoring SKILL.md metadata, improving discoverability, or registry index entries. Matches hub tokenized search on slug and description.

  chain add skill-search-authoring

  Source

• team-hub-standards

  New Protected Since v0.3.0
  • automation

  Shared team hub conventions: global.md, Git sync, registry buckets, and conflict policy. Use for team hubs; documents current behavior and roadmap team bucket. Not a substitute for chain sync setup.

  chain add team-hub-standards

  Source

• update-cli-config

  Authoring Since v0.3.0
  • automation

  Route terminal/agent CLI configuration to the correct product and file, then edit safely. Covers Cursor CLI (~/.cursor/cli-config.json) in full detail. Use when the user asks about CLI permissions, approval mode, vim mode in a terminal agent, sandbox, or agent CLI display options — after confirming which CLI they use. For GUI editor preferences (themes, format on save), use update-editor-settings instead.

  chain add update-cli-config

  Source

• update-editor-settings

  Authoring Since v0.3.0
  • web
  • automation

  Modify editor settings for VS Code–family editors (settings.json: VS Code, Cursor, VSCodium) or get guidance for JetBrains, Zed, and Neovim. Use when the user wants to change editor settings, preferences, themes, font size, tab size, format on save, auto save, keybindings, or language-specific options.

  chain add update-editor-settings

  Source

No skills match your search or filter.