mattpocock-skills
claudeMatt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Planner → architect → coder(+navigator⇄driver) → code-reviewer pipeline. /plan decomposes a brief into epics, stories and ARCH handoffs; /build cascades an epic onto one branch, one squash commit per story, building each story SOLO or as a navigator⇄driver TDD pair, and opens a PR. Handoffs, a tooling-debt ledger, ping-pong pairing and model tiering are part of the protocol. /review spawns the code-reviewer against an open PR, posting one round comment with stable findings that the coder's REVISE mode closes. --fast on either command takes the lean lane instead: a flat task list, decide-and-log instead of handoffs, only the tests that catch expensive bugs, and every shortcut in the ledger.
Plugin agentic-sdlc from marketplace sanimara in tofbabs/agent-sdlc. Commands read from the repo's manifest — 17 days ago.
/agentic-sdlc:buildBuild an epic to completion — stories cascade onto one epic branch, each built SOLO or as a navigator⇄driver TDD pair and landed as ONE squash commit. Architect unblocks mid-build. Opens one PR. Review is `/agentic-sdlc:review`; the REVISE loop that closes its findings runs here separately. `--fast`
/agentic-sdlc:planPlan a feature — stories from the planner, architecture decisions from the architect. Non-blocking. `--fast` takes the lean lane: a flat task list, no handoffs.
/agentic-sdlc:reviewReview an open PR — the code-reviewer reads it in its own worktree and posts one "Review — round <k>" comment with findings F1..Fn, then the GitHub verdict. Never edits, never merges. Closing its findings is /build's REVISE loop.
/agentic-sdlc:waveRun a wave of brief Issues; `start` boards them, `status` reads the board.
architectResolves architecture and tooling decisions. Handles handoffs from the planner and mid-build blocks from the coder. Picks tools, patterns, and structure. Writes lightweight ADRs for decisions that are expensive to reverse.
code-reviewerReviews one open pull request against its story, the ACCEPTED ADRs, the tooling-debt ledger and the surrounding code, and posts ONE "Review — round <k>" comment with a verdict and stable finding IDs, then the GitHub verdict. Never edits code, never merges. Spawned by /review; its findings are closed
coderImplements a story SOLO (small stories, end to end), in FAST mode (a lean task with one observable check), or in REVISE mode addressing findings from the code-reviewer. (A PAIR story's driver is the separate `driver` agent.) Blocks to the architect on unanticipated decisions. Logs tooling gaps it de
driverThe driver half of a PAIR-mode TDD loop. Makes the navigator's failing tests for one behaviour pass with the simplest implementation, commits, logs, stops. Never writes tests. A fresh agent every alternation, spawned by /build's PAIR LOOP.
navigatorThe navigator half of a pair-programming loop. Writes the failing tests for the next behaviour, reviews the driver's last increment, steers, and closes the story once every AC is green. Never writes implementation code. Alternates with the driver, one increment at a time. Used by /build for stories
plannerDecomposes a brief into epics and stories with acceptance criteria, or in FAST mode into a flat task list with one observable check each. Flags decisions it should not make itself as ARCH handoffs for the architect. Use PROACTIVELY at the start of any feature.
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.