mattpocock-skills
claudeMatt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Route the task, explore or design before code, plan in phases with a matching loop prompt, work one slice at a time, review on the running app, hand off.
Plugin flow from marketplace myworkflow in KhaiStimpson/myworkflow. Commands read from the repo's manifest — 1 month 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.
designAnswer an under-determined visual question with three variants in one self-contained HTML artifact, at desktop and mobile, then record the decision so implementation reuses the real design system. Use before writing UI code whenever more than one defensible visual answer exists.
exploreBrainstorm an open-ended idea without touching a single file, and land it as an artifact of concrete options with a recommendation. Use when the request is "brainstorm", "come up with N suggestions", "be creative", "what could we do here" — anything where the answer is ideas, not code.
eyesRun the app, look at it at desktop and mobile widths, and capture durable screenshots — the review pass that green tests do not perform. Use before committing any UI-touching change, and before asking for approval on visual work.
flowFront door for any new piece of work — routes it, then runs the route end to end. Use when you want the whole working loop rather than one stage of it.
loopStart the loop that executes the live plan — read the plan, fill in the loop prompt from the template, and hand it to /loop. Use to run a planned effort to completion or to resume one, instead of pasting the loop prompt by hand.
planTurn agreed work into a phased markdown plan with checkboxes in docs/, plus the exact loop prompt that executes it — and implement none of it. Use when work exceeds one reviewable change, or when asked for "a plan with phases", "a loop command", or "a goal prompt".
startThe front door for any new piece of work. Investigate the task, announce a route in three lines with a one-token veto, then run it. Use when work arrives and it is not obvious whether it needs a plan, a design pass, a single slice, or just doing.
workImplement exactly one slice — one task or one ticket — on its own branch, verify it, capture screenshots if it touches UI, and land it on the integration branch. Use as the body of a loop iteration, or for a single self-contained change.
wrapClose out a session or an effort — write the handoff file, open the integration PR, comment the tracker, and make the repo present itself. Use when stopping mid-effort, when a loop finishes, or when the work is done.
fresh-eyesReview one landed slice against the standards it claims to follow and the task it claims to complete, with no memory of writing it. Use after a slice is verified and before it is merged, or on any diff that wants a second read.
researcherBackground research on any load-bearing external fact — vendor terms, API limits, platform behaviour, pricing, licensing, framework guidance. Writes a cited report to docs/research/ and returns a summary. Use whenever a decision rests on something outside the repo, instead of asserting it from memor
"${CLAUDE_PLUGIN_ROOT}"/scripts/session-state.sh
"${CLAUDE_PLUGIN_ROOT}"/scripts/guard-image-read.sh
"${CLAUDE_PLUGIN_ROOT}"/scripts/phase-boundary.sh
"${CLAUDE_PLUGIN_ROOT}"/scripts/context-budget.sh
"${CLAUDE_PLUGIN_ROOT}"/scripts/write-handoff-spine.sh
"${CLAUDE_PLUGIN_ROOT}"/scripts/write-handoff-spine.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
Backend API design, GraphQL architecture, workflow orchestration with Temporal, and test-driven backend development
9 skills · 1 command · 8 agents
Long-term semantic memory for Claude Code, powered by OpenViking
5 skills · 1 command · 1 MCP server · hooks
AI-supervised issue tracker for coding workflows
1 skill · 1 agent · hooks
PRD generation and conversion skills for the Ralph autonomous agent loop
2 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.