workflow-toolkit

[DEPRECATED] Bootstrap a new project with Claude Code best practices - sets up context management, skills directory, and initial documentation

Check dependency health before adding to project - verifies npm/CDN versions, bundle size, maintenance status, and known issues

Clean up git branches - removes merged branches, warns about stale branches, syncs with remote

Safe commit workflow - runs build, tests, and checks before allowing commit

Orchestrate a comprehensive post-implementation review with multiple phases (architecture, code, tests, security, docs)

Staff-level Rust code review specialist. Use PROACTIVELY after significant code changes for security, reliability, and accessibility reviews. Expert in planning implementations and documenting changes.

Verify that operations are idempotent by running them twice and comparing results. Use for sync operations, data migrations, API calls, or any operation that should produce the same outcome when repeated.

Generate operational documentation (troubleshooting, performance, deployment) by analyzing actual codebase patterns. Use when you need to create or update ops docs that reflect real implementation.

Coordinate multi-phase code reviews by delegating to specialized review agents and managing the branch/PR workflow. Use when conducting comprehensive reviews that require architecture, code quality, test coverage, and security analysis.

Three-layer Stop hook that detects and blocks work-skipping rationalizations by Claude Code agents. Prevents agents from unilaterally deciding to skip assigned work, bypass mandatory processes, or cite unverified context constraints as justification for shortcuts. Layer 1: Deterministic regex detection of known laziness phrases (Tiers 1-2) Layer 2: Haiku-evaluated intent detection for premature victory and silent omission (Tier 4) Layer 3: Optional agent-based deep verification against task lists and plan files This is a passive hook — it activates automatically when the plugin is enabled. No slash command needed. The hook fires on every Stop event and only blocks when laziness patterns are detected.

GitHub Agentic Workflows (gh-aw) — write AI-powered automation workflows in natural-language markdown that compile to secure GitHub Actions. Supports Copilot, Claude, and Codex engines with safe-output guardrails, MCP tool integration, and sandboxed execution. Use this skill when users need to: (1) Install or set up gh-aw in a repository (2) Create, edit, or compile agentic workflow markdown files (3) Configure triggers, schedules, safe outputs, tools, or MCP servers (4) Switch AI engines (Copilot, Claude, Codex) or configure engine options (5) Troubleshoot workflow failures, compilation errors, or permission issues (6) Understand gh-aw patterns (ChatOps, DailyOps, IssueOps, etc.) (7) Configure security: permissions, network rules, sandbox, threat detection (8) Use advanced features: memory, imports, orchestration, cross-repo ops

Prepare Claude Code skills, agents, or collections for open-source sharing on GitHub. Supports two modes: (1) Standalone repo creation, or (2) Marketplace integration into existing plugin repos. Use when: - "prepare for open source", "open source this skill" - "upload skill to github", "share this agent" - "add to marketplace", "add to robot-tools" - "create repo for skill", "package for sharing" - User has a skill directory, agent file, or collection to share

Multi-stage implementation review with parallel sub-agents, severity-based autonomous fixes, and gated test verification. Runs code quality, architecture, simplicity, documentation, and security reviews in sequence with test gates between each fix stage. Security review is blocked until all other fixes are complete. Use after completing a feature, implementation phase, or release candidate. Supports scope modes: full, code-only, security, simplicity, docs.

Validates plugin manifest consistency, README cross-references, SKILL.md frontmatter, and version sync across the robot-tools monorepo. Has two modes: - Validate mode: checks and reports pass/warn/fail (default) - Release prep mode: validates, walks through version bumping, re-validates, summarizes for commit Claude should proactively suggest release-prep mode when a development session that added or modified plugin content appears to be wrapping up.

3 hooks across 1 event