mattpocock-skills
claudeMatt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Coordinator plugin: plan-enrich-chunk-review pipeline, named reviewer personas (the Staff Engineer, the Director of Engineering, the Game Dev Reviewer, the Data Science Reviewer, the UX Reviewer, the Front-End Reviewer), fan-out dispatch, session continuity via handoffs and flight recorders, and workflow skills (/plan, /review, /execute-plan, /workday-start, /workstream-complete).
Plugin coordinator from marketplace coordinator-claude in dbc-oduffy/coordinator-claude. Commands read from the repo's manifest — 7 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.
app-sessionCensus/launch/teardown/validate for a repo's declared app.
architecture-auditRotational arch audit: score systems, audit the top. No code edits.
artifact-investigationOpaque binary or artifact investigation: structure first, script second, agent last.
box-onboardingAdmit a new macOS box to the multibox fleet: Homebrew python3, Remote Login, install, engine, keys, Unimatrix, home repo, verify.
brainstormingShape vague or multi-subsystem asks into requirements. Size first if unsized.
bug-sweepCodebase bug hunt: fix AI-fixable bugs, defer the rest.
cloud-channelcloud-spawncode-healthconsolidate-gitcoordinator-updatedebt-triagedesign-canvasdogfoodenrich-and-revieweval-outputexecute-planfinishing-a-development-branchgoal-blitzgoal-settinggroup-emhandofflearn-lessonsmerging-to-mainnew-projectparallel-code-reviewpickupplan-blitzplan-delivery-auditplanquick-wraprepo-setupreviewroadmap-blitzroadmap-planningsetupshapesizingspikespinoffstaff-sessionstrategic-self-description-refreshsystematic-debugginguhuravalidateworkstream-completeworkstream-start/coordinator:architecture-surveyBuild or refresh the architecture atlas.
/coordinator:autonomousToggle autonomous mode — the PM-away premise for what counts as a legitimate stop.
/coordinator:bug-blitzSubmit a bug-blitz as one coordinator:bug-blitz Workflow call.
/coordinator:cruft-sweepFind reclaimable scratch and orphans; apply only if confirmed.
/coordinator:distillDistill session artifacts to wiki and decisions; archive specs, drop scratch.
/coordinator:fire-and-forget\"fire and forget\" / \"fire and forget\": a raw ask or sizing to a reviewed commit in one fire.
/coordinator:installInstalls the coordinator plugin: prereqs, project config.
/coordinator:mise-en-placeAutonomous backlog run — flight-recorder prep, then run-through.
/coordinator:mise-prepCertify approved plans; route residue to plan-author-residue.
/coordinator:notebooklm-researchPM-GATED, never from a subagent. NotebookLM research for video/audio sources.
/coordinator:researchResearch by value class: scouts, corpus, deep, or spike. EM-fired, never from a subagent.
/coordinator:review-codeReview a code diff/PR, or apply landed findings — findings land on the artifact.
/coordinator:uninstallReverses the coordinator install, fully or back to marketplace.
/coordinator:update-docsSync all documentation artifacts to the current codebase state.
/coordinator:workday-completeEnd-of-day wrap — validate, consolidate, review, changelog.
/coordinator:workday-startMorning orient — triage handoffs, staleness, priorities.
/coordinator:workweek-completeWeekly release ceremony — validate, docs, release notes, merge.
/coordinator:workweek-startWeekly orient — review last week, set this week's priorities.
apmPersonas are Opus-only. Angelique, adversarial junior-PM: ELI5 a plan's choice, then challenge it.
atlas-clarity-reviewerClarity verdict for one atlas page: followability, ambiguity. Never accuracy.
blitz-emPersonas are Opus-only. EM judgment in a plan-blitz wave: interrogates sizings.
code-reviewer-weeklyStatic weekly-gate chunk reviewer, compaction-safe. Only via parallel-code-review.
code-reviewerStatic Sonnet code reviewer: nits, weak tests, security. Sidecar + OK/WARN/BLOCKED.
coverage-auditorFresh-eyes check of specialist claims against a finished synthesis: present/absent.
delivery-verifierChecks executor delivery claims against the frozen diff. Read-only.
dep-cve-auditorAudits dependency manifests for CVEs; source is security-audit-worker's.
distill-writerLean batch writer for distill-harvest: synthesizes a batch of topic clusters into staged files.
doc-link-checkerValidates doc links and anchors; external URLs only on opt-in. Returns a status table.
docs-checkerVerifies external API claims against authoritative docs before Opus review. A table.
eng-directorPersonas are Opus-only. The Director of Engineering, Director of Engineering: staff rigor, cross-repo boundaries.
enricherEnriches plan stubs and live plan bodies pre-execution. Verifies facts, never decides.
executorDelivers one chunk of an enriched, reviewed plan; validates at chunk boundaries.
exit-criterion-falsifierAuthors and runs the baseline falsifier for a plan's prime exit criterion.
exit-criterion-judgeTerminal judge: is a run's prime exit criterion met or not_met.
external-pattern-checkerOpt-in web research on plan claims prior-art-checker left Silent.
external-reviewerOpt-in non-Anthropic review lens: runs the codex CLI out of process, applies findings, owns the ledger.
falsifier-integrity-reviewerStatic review of an exit-criterion falsifier: can it report red at all.
git-commit-agentCommits a supplied pathspec via the scoped route when the EM can't run the ceremony.
group-em-assistantSonnet aide to a Group EM: memo inbox, park-spool triage. Read-and-report.
notebooklm-research-scoutHaiku NotebookLM scout: finds sources per notebook topic, writes sources.md.
objective-reacherWrites a delta plan whose rows target only the failing falsifier axes of a run objective; never declares done.
op-relayRelays one coordinator-invoke op verbatim.
overengineering-reviewerPersonas are Opus-only. Kira: final whole-diff pass; cuts waste, may change code, checks the target. Not plans.
parallel-review-synthesizerFolds chunk reviews and specialists into one verdict; parallel-code-review only.
plan-author-residueAmends plans in place with /mise-prep capabilities and appends; never authors one.
plan-authorAuthors one plan doc against a finalised size/route; never sizes, routes, or executes.
plan-coverage-checkerMechanical plan check: slate vs oracle, deferral ratification, task-spine. No fixes.
plan-reviewerSonnet plan-body reviewer: assumes gaps, runs the coverage checklist; edits only as XS/S reviewer.
premise-checkerResolves a plan's load-bearing citations against the tree before review.
prior-art-checkerRecall pre-flight: plan vs wikis, decisions, lessons. Conflicts/Compatible/Silent.
queue-grind-op-runnerRuns one queue-grind op verify via coordinator-invoke; relays exit code and JSON.
repo-scoutHaiku repo-research scout: inventories assigned repo chunks. Mechanical only.
repo-specialistSonnet repo-research specialist: deep-reads a scout inventory, writes claims.json.
research-brief-authorSonnet research-brief author: after the haiku source survey, writes the expert brief specialists read.
research-scoutHaiku web-research scout: runs scope.md queries, writes the source corpus.
research-specialistSonnet web-research specialist: deep-reads scout sources, writes claims.json.
research-sweepOpus NotebookLM sweep after workers: coverage check, gap fill, final framing.
research-synthesizerOpus web-research sweep after specialists: adversarial coverage, gaps, summary.
research-workerSonnet NotebookLM worker: ingests scout sources, queries, writes {letter}-claims.json.
residue-authorAuthors /mise-prep residue: capabilities, appends, census rewrites; never decides scope.
roadmap-plannerAuthors a roadmap's inventory, clusters, overview, stubs; never sizes, approves, executes.
security-audit-workerSonnet static-security scan of a diff: traversal, injection, secrets.
senior-front-endPersonas are Opus-only. The Front-End Reviewer: front-end review for design-system adherence.
sizing-scoutSizes one ask by engineering depth, not breadth; in engine-ask it also writes and gates it.
spike-workerSonnet spike worker: web research plus scratch-dir probes to a binary viable/not-viable verdict record.
staff-data-sciPersonas are Opus-only. The Data Science Reviewer: data science, ML, and statistics review.
staff-engPersonas are Opus-only. The Staff Engineer: generalist staff-engineer review of code, plans, docs.
staff-uxPersonas are Opus-only. The UX Reviewer: user-facing flow review for clarity and trust.
structured-synthesizerOpus structured-research synthesizer: verifier findings to schema YAML/JSON.
subtractive-adjudicatorPersonas are Opus-only. Terminal subtractive pass: what comes OUT of reviewer additions.
test-evidence-parserClassifies captured test failures (real/flake/env/timeout/skip). Mechanical.
test-runnerRuns a diff's scoped tests and execute-review prep. Never the fast/full suite.
vp-productPersonas are Opus-only. The VP-Product Reviewer, VP Product: stress-tests choices before they ship.
workflow-makerEmits a plan's spine as a Workflow script; returns path and wave map for the EM to fire.
python3
python3
python3
python3
python3
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" hooks.guard_manufactured_blocker 2>/dev/null; rc=$?; [ $rc -eq 2 ] && rc=0; exit $rc
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" hooks.guard_turn_intent 2>/dev/null; rc=$?; [ $rc -eq 2 ] && rc=0; exit $rc
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" --advisory hooks.sessionend_archive_session 2>/dev/null; rc=$?; [ $rc -eq 2 ] && rc=0;
python3
python3
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" --advisory hooks.workflow_step_dispatch 2>/dev/null; rc=$?; [ $rc -eq 2 ] && rc=0; exit
python3
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" --advisory hooks.workflow_step_dispatch 2>/dev/null; rc=$?; [ $rc -eq 2 ] && rc=0; exit
python3
python3
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" hooks.preuse_bash_dispatch
python3
python3
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" hooks.guard_agent_route_admission
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" --advisory hooks.guard_ue_specialist_advisory
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" hooks.guard_brief_prescribed_method
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" hooks.guard_editor_python_direct
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" --advisory hooks.preuse_search_dispatch 2>/dev/null; rc=$?; [ $rc -eq 2 ] && rc=0; exit
python3
python3
python3
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" --advisory hooks.block_group_em_agent_cap_grant 2>/dev/null; rc=$?; [ $rc -eq 2 ] && rc
COORDINATOR_DOOR_STDIN_MODE=hook "${COORDINATOR_SETTINGS_HOME:-$HOME/.coordinator-claude-settings}/bin/hook-run" --advisory hooks.rewrite_sendmessage_peer_name 2>/dev/null; rc=$?; [ $rc -eq 2 ] && rc=
python3
python3
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.