Routes you to the right mattpocock skill or flow for your current situation.

Pursue a long-running goal in one session by coordinating multiple subagents.

Hand the current conversation to a fresh background agent that continues the work immediately.

Reviews changes against repo coding standards and the originating spec, using parallel sub-agents.

Gives a shared deep-module vocabulary for designing interfaces, seams, and more testable code.

Runs a structured loop to diagnose hard bugs, throwing errors, and performance regressions.

Helps build and sharpen domain terminology, glossaries, and architecture decision records.

Block destructive git commands like push, reset --hard, and branch -D via Claude Code hooks.

Get relentlessly interviewed to expose gaps and sharpen a plan or design.

Stress-test a plan or design through relentless questioning while capturing ADRs and glossary terms.

Stress-test your plan, decision, or idea with relentless pointed questioning.

Compact the current conversation into a handoff document for another agent to resume.

Build a piece of work in code from a spec or ticket set.

Implement code from /to-spec and /to-tickets outputs.

Find codebase deepening opportunities, review them in an HTML report, then refine one.

Interview me about specs for workflows I want to build in this workspace.

Replace unsafe as type assertions in tests with Shoehorn for partial test data.

Write a pull request body that describes changes clearly for reviewers.

Build a throwaway prototype to test a state model or explore UI design.

Investigate questions in primary sources and save findings as a Markdown file.

Review a coding session to capture lessons and improve future work.

Generate lint-passing exercise directories with sections, problems, solutions, and explainers.

Configure a repo's issue tracker, triage labels, and domain docs for engineering skills.

Add Husky pre-commit hooks running Prettier, type checks, and tests on commit.

Configure dependency-cruiser so TypeScript packages expose only their entry-point modules.

Build features or fix bugs test-first using red-green-refactor cycles.

Teach you a new skill or concept directly inside your current workspace.

Turn an unanswerable decision into a questionnaire for someone else to complete.

Turn the current conversation into a published spec by synthesizing it, without interviewing.

Split a plan or conversation into tracer-bullet tickets with blocking edges, published to a tracker.

Route issues and external PRs through triage states and write agent-ready briefs.

Make the agent re-pitch a message that failed to land clearly.

Map large multi-session work as decision tickets and resolve them until the path is clear.

Generate an interactive bash wizard that guides humans through steps only they can perform.

Assemble raw material into a beat-by-beat narrative, grounding terms before relying on them.

Write effective skills, AGENTS.md, and CLAUDE.md files that agents follow reliably.

Mine raw fragments of ideas before any structure exists.

Turn raw material into a finished article, shaping it paragraph by paragraph.