mattpocock-skills
claudeMatt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
An agentic dev team installed per project: scout, planner, tester, coder, auditor (Claude or Codex), releaser and consultant, each with its own context, tools, model and effort, customised to the repo through a reviewed profile. Among subagents, only the releaser may push
Plugin dev-team from marketplace xiaolai in xiaolai/dev-team-for-claude. 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.
auditAudits the current change, or a plan, with dev-team's auditor in refute mode, choosing Claude or Codex as the engine. Use when the user wants an independent review of uncommitted work, of a branch against a base commit, or of a plan, outside a full /dev-team:run.
initSurveys this repository and drafts its dev-team profile, `.claude/dev-team.md`, which every dev-team agent reads first - the check commands (verified by running them), conventions, protected paths, release steps and the auditor's default engine. Use when setting up dev-team in a repository, or when
runRuns the dev-team on a task - survey for larger tasks, plan, tests first, implementation, an independent audit and a bounded fix loop - with every handoff saved under `.dev-team/`. Use when the user wants a feature, fix or change built by the team rather than inline.
_auditorAudits a change, or a plan before any code is written, in refute mode, with Claude or Codex as the auditing engine: `engine: claude` or `engine: codex` in the request picks it, and the repository profile sets the default. Every finding names the check that would expose it. Runs on Opus at xhigh effo
_coderImplements a planned change until the tester's tests and the profile's full gate pass, without weakening any test. Runs on Opus at medium effort. Not for writing the tests that decide whether the change is done (use _tester), reviewing the change (use _auditor), or committing and pushing (use _relea
_consultantAnswers the open questions any role hits mid-task, such as which library to use, what a spec or vendor document actually says, or how an approach fails in practice. It researches primary sources first and recommends second. Read-only with web access; runs on Opus at xhigh effort. Not for surveying t
_plannerTurns a task into success criteria that can each be checked by a command or a test, then a step-by-step plan, working from a scout brief and the repository profile. Read-only; runs on Opus at xhigh effort. Not for writing code or tests (use _tester and _coder), or for questions that need sources out
_releaserCommits, versions and publishes a finished change by following the repository profile's release steps, and is the only subagent the plugin lets push or publish. Pushes only when the request says the user approved it. Runs on Sonnet at low effort. Not for code changes beyond version bumps and release
_scoutSurveys the repository before planning: the files a task touches, how tests and builds run, the conventions in force, and the history that explains the current code; or, for /dev-team:init, the facts a repository profile needs. Read-only, runs on Sonnet at low effort, and returns a brief with file:l
_testerWrites the tests for a planned change from the plan's success criteria, before the implementation exists, and confirms they fail for the right reason. Runs on Opus at high effort. Not for implementing the change (use _coder), reviewing it (use _auditor), or committing (use _releaser). <example> Cont
node "${CLAUDE_PLUGIN_ROOT}/scripts/push-guard.mjs"
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.