mattpocock-skills
claudeMatt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Subagents let one session use several model tiers at once. This plugin makes that deliberate: a PreToolUse gate blocks any Workflow whose agent() calls have no model/effort, then, in manual mode, asks you to approve the routing plan, offering cheaper AND higher-quality directions (auto mode, the default, decides and lets it run). A finish check warns when a completed Workflow's journal shows agents that never returned. Per-project routing rules live in agenting/AGENTING.md; mode, disposition and a workflows opt-in persist per chat across compaction. Also a mid-session cache-cost tripwire, a tokenomics script, 20 tiered agent definitions and an end-to-end health check. A fan-out view shows clean runs (cleanrun.py, clean_runtime.mjs) the way Workflows are shown: a row above the prompt and a /fanouts pane.
Not in a marketplace yet, so it loads for one session with --plugin-dir.
Plugin agenting in furkanzt/claude-agenting. Commands read from the repo's manifest — 3 days ago.
This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.
agentingUse when writing or editing a Workflow script, planning a fan-out of many agents, choosing a model or effort for an agent() or Agent call, when the user switches agenting mode (auto/manual) or disposition (fast/balanced/quality) or opts in to workflows for this chat, or when a project has agenting/A
/agenting:agenting-checkVerify agent routing is wired up and actually working (end-to-end, not config-reading)
/agenting:agenting-modeReport or set this chat's agenting mode (auto/manual), disposition (fast/balanced/quality) and workflows opt-in (workflows on/off)
cicd-engineerDevOps and deployment pipeline expert. Use for GitHub Actions, Docker, PM2, environment management, rollback planning.
coder-tsTypeScript and modern frontend/backend development specialist. Strict mode, generic types, React/Next.js, Fastify, Zod. Use for complex type design and architectural coding.
coderGeneral-purpose implementation agent for routine code - boilerplate, CRUD endpoints, migrations, config files, simple components, repetitive edits across files. Use when the pattern is already established and the work is applying it. Prefer coder-ts for complex TypeScript type design and architectur
coordinatorMUST BE USED PROACTIVELY for any task involving 3+ files, multi-area development (backend+frontend), full feature implementation, or large refactoring. Creates Agent Teams, orchestrates other agents, and routes each one to the cheapest tier that can do the job.
database-architectDatabase design and query optimization specialist. Use for schema design, index strategy, migration management, N+1 detection.
frontend-devUI/UX implementation and CSS/styling expert. Use for React, Vue, Svelte, Tailwind, accessibility, Core Web Vitals optimization.
git-specialistVersion control strategy and Git workflow expert. Use for branching strategy, conventional commits, merge conflict resolution, monorepo management.
incident-responderFast triage and resolution expert for production issues. Use for P0/P1/P2/P3 severity classification, root cause analysis, rollback planning.
ml-developerMachine learning and AI integration specialist. Use for Anthropic SDK, LangChain, LlamaIndex, embedding, vector search, RAG pipeline design.
performance-engineerPerformance optimization and profiling specialist. Use for bundle size, query optimization, N+1 detection, benchmark analysis.
refactoring-specialistTechnical debt analysis and incremental refactoring specialist. Use for SOLID, DRY, YAGNI, code smell detection, legacy code modernization.
researcherCodebase and document research agent. Use for reading files, tracing how something works, mapping dependencies, finding where a symbol or pattern lives, and reporting back a compact summary instead of raw file contents. Prefer scanner for pure counting and inventory; prefer this when the answer requ
reviewerReviews code quality, security, and best practices. Used for quality-critical work such as PR review, large-change checks, refactoring approval. Does NOT delegate to another agent — uses its own reasoning power.
scannerRead-only inventory and counting agent. Use for mechanical corpus reconnaissance where no judgment is required - listing files, counting pages or lines, measuring sizes, collecting grep hits, tallying matches, reporting directory structure. Do NOT use for analysis, review, or any task where the answ
security-architectSecurity analysis and OWASP compliance specialist. Use for PR review, auth implementation, secrets audit, dependency vulnerability scanning.
swarm-coordinatorOrchestrator that distributes complex tasks across parallel agent swarms. Fully supports the 9-layer delegation system.
system-architectSpecialist agent for high-level system design and architectural decisions. Use for microservice/monolith trade-offs, API design, database selection, scalability decisions.
tech-writerDocumentation agent for README files, CLAUDE.md updates, changelog entries, ADRs, and API docs. Use when prose needs to be written or updated to match an existing document's voice and structure. Not for code comments inside implementation work - the implementing agent writes those.
tester-tddTest-Driven Development and quality assurance expert. London School TDD approach. Use for writing tests, coverage analysis, flaky test detection.
text-mechanicMechanical text transformation agent. Use for cleanup where the output must be a rearrangement of the input and nothing may be invented - de-hyphenating line-wrapped words, reflowing broken paragraphs, stripping repeated page headers and footers, normalizing whitespace, converting an already-identif
python3 "${CLAUDE_PLUGIN_ROOT}/hooks/workflow-routing-guard.py"
python3 "${CLAUDE_PLUGIN_ROOT}/hooks/cache-tripwire.py"
python3 "${CLAUDE_PLUGIN_ROOT}/hooks/workflow-finish-check.py"
python3 "${CLAUDE_PLUGIN_ROOT}/hooks/session-continuity.py"
Problems found in this repo's manifests
Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
15 skills · hooks
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
Cross-session memory and token savings for coding agents.
6 skills · 1 agent · 1 MCP server · hooks
Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.
1 MCP server
Automates browser interactions for web testing, form filling, screenshots, and data extraction
12 skills
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.