mattpocock-skills
claudeMatt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Thirteen skills for the whole development lifecycle — writing and de-AI-ing code, review, testing, release, and docs — built through the skill-creator six-gate pipeline, plus hook-enforced session memory and branch protection.
Plugin dev-workflow-pack from marketplace dev-workflow-pack-local in rolling-codes/AI-Coding-Swiss-Army-Knife. Commands read from the repo's manifest — 24 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.
ai-hygieneAI-fingerprint hygiene for code and prose. Prevent AI tells when writing new code or adding a dependency (reuse, simplicity, deps); remove them from existing code or prose; handle watermarks. Use when the user says "keep it simple", "sound like AI?", "humanize this", "audit this code", "remove AI pa
architecture-reviewUse this to assess structural health of a codebase or a proposed change — coupling, cohesion, layering violations, dependency direction, future maintenance cost — when asked "review the architecture", "is this well structured", "should this be split up", "check for layering violations", or before a
bug-triageUse this to consolidate, deduplicate, and severity-rank 2 or more existing bug findings into one actionable report — "triage these", "which bugs should I fix first", "consolidate findings", or a dump of multiple reports or review-agent output. Use code-review to find new bugs in a diff; use this whe
changelogUse this to generate or update CHANGELOG.md from git commit history in Keep a Changelog format when the user wants changes documented — "update changelog", "release notes", "what changed since last release"; NOT for single commit messages (commit-message), NOT for PR bodies (pr-description), and NOT
code-reviewUse this to review code changes before merge — self-review of your own diff, or dispatching the code-reviewer subagent for an independent pass on larger work — when the user wants changes checked before pushing or merging: "review my changes", "is this ready to merge", "self-review", "check my diff"
commit-messageUse this to generate a Conventional Commits-compliant message from staged git changes when changes are staged and a message is needed — "write a commit message", "help me commit", "what should I commit this as"; NOT for PR descriptions or changelog entries (pr-description / changelog), and NOT for d
context-compressionUse this to manage token budget explicitly during long sessions — summarize, decide keep vs. drop, and age out stale context — when /cost shows context ≥ 50% (or the context bar is past halfway in the IDE), before a deliberate manual /compact, or on "context is getting long", "summarize what we've d
dependency-checkUse this to check whether dependencies are up to date and flag outdated, deprecated, or license-incompatible packages — "are our dependencies up to date", "check for outdated packages", "any deprecated packages", "license compliance check"; NOT for vulnerability scanning (security-audit), NOT for fi
dev-workflowUse when the user gives an everyday dev task or asks what to work on next, to open an issue, branch, push, check CI, or merge — runs the Research→Plan→TDD→Review→Commit pipeline, GitHub ops, session memory, and model routing. Trigger phrases: "fix this bug", "implement X", "add a feature", "refactor
docs-auditUse this to audit project documentation (README, CONTRIBUTING, API docs, wikis, in-repo .md files) against what the codebase actually does now, and prune or update what's stale, when asked "review the docs", "are the docs out of date", "clean up the README", "docs don't match the code anymore", or a
infra-reviewUse when the user asks to review IaC files — "review this Dockerfile", "is this k8s config production-ready", "check the terraform", "audit the CI workflow", "is this GitHub Actions safe"; NOT for application code correctness (code-review), NOT for dependency vulnerabilities (security-audit), and NO
kill-testUse when the user needs a pre-build go/no-go gate before implementing something — "should I build this", "is this worth implementing", "kill test this idea", "go/no-go on this feature", "should I write this from scratch or is there already something"; NOT for reviewing finished code (code-review), N
observability-auditUse this to audit logging, tracing, and error-handling coverage across a codebase — "do we have enough logging", "add observability", "check our error handling coverage", "are we tracing this", "alerting gaps", "what would we miss if this broke in prod"; NOT for adding new log statements (dev-workfl
performance-auditUse when the user asks about slow code, bottlenecks, or profiling — "why is this slow", "profile this function", "what's the hot path", "benchmark before/after", "where is time being spent"; NOT for writing the optimization code (dev-workflow), NOT for performance test cases (test-strategy), and NOT
pr-descriptionUse this to generate a structured pull request description from the current branch's diff and commit history when the user is opening a PR or asks what it should say — "write a PR", "draft a pull request", "PR description", "opening a merge request"; NOT for single-commit messages (commit-message),
refactor-guideUse when a refactor spans multiple files or involves a framework migration — "help me migrate from X to Y", "break up this monolith", "refactor this module safely", "extract this into a service"; NOT for small local refactors during active coding (dev-workflow), NOT for reviewing a finished diff (co
release-prepUse this to validate a release before it ships — detect version drift across all sources, verify the changelog, confirm tests pass, and produce a go/no-go checklist — when the user is about to tag or publish: "ready to release", "pre-release check", "can I ship this", "version drift", "about to tag"
scope-creepUse this to detect and assess scope creep when a new feature, change, or dependency is introduced mid-build — fires automatically during dev-workflow and code-review sessions without being invoked, and can also be invoked directly on "while we're at it", "can we also", "one more thing", "quickly add
security-auditUse this to run a security audit on a repository — dependency vulnerability scanning, SAST, and secrets detection — when the user asks "check for vulnerabilities", "scan for CVEs", "audit dependencies", "any secrets in this code", "OWASP check", or "is this safe to ship"; NOT for general code qualit
strategic-compactSuggests manual context compaction at logical intervals to preserve context through task phases rather than arbitrary auto-compaction. Use when a session is approaching a context limit and a task phase is a natural place to compact.
test-strategyUse this to generate a test plan and the tests themselves — unit, integration, edge cases, and regression tests — when asked "write tests for this", "what edge cases am I missing", "add a regression test", "test coverage for this change", or right after a bug fix before it's considered done; NOT for
code-reviewerIndependent read-only review of a git commit range against its requirements. Dispatch after completing a major feature or before merge, with a description of what was built, the requirements or plan, and the BASE and HEAD commit SHAs.
sh "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/protect-branches.sh"
sh "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/guard-destructive.sh"
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
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
Build powerful, stateful CLI interfaces for any GUI application using the cli-anything harness methodology.
5 commands
Automates browser interactions for web testing, form filling, screenshots, and data extraction
12 skills
Perform AI code review on Git diffs — supports workspace changes, branch ranges, and single commits with concurrent per-file analysis, codebase search, and deep context-aware review.
2 commands
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.