mattpocock-skills
claudeMatt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
The six-stage AI-native SDLC playbook as commands, skills and opt-in hooks: intent, design with Gherkin scenarios and red bindings, plan then build until green, a fresh verifier, a Copilot-reviewed pull request, and a deterministic CI-drift watcher. Plus a reflect step that turns a repo's corrections into the next version. Project-agnostic; every artefact is plain markdown in the project's own git.
Not in a marketplace yet, so it loads for one session with --plugin-dir.
Plugin balka in meteoFurletov/balka. Commands read from the repo's manifest — yesterday.
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.
balka-artifactsWrite or update Balka SDLC artifacts in an opted-in repository when no /balka command was typed. Covers intent, spec, plan, Gherkin, estate facts, and proposals; excludes ordinary documentation outside the loop.
balka-watch-writebackUse when a Stage 6 CI drift detection has fired and the diagnosis must be written back as a draft intent.md — normally a headless `claude -p` run started by the balka-watch GitHub Actions workflow, whose prompt carries a band breach, a rolling baseline mean and sigma, and a daily failure-rate series
/balka:boardShow the board — every change in this repo, across every worktree, as a kanban with each card's next command. Opens live in the Claude desktop app's Browser pane; prints as markdown anywhere else.
/balka:buildStage 3, Build, second half — implement an accepted plan.md and run the verify commands until they are green, without stopping for anything the plan already settled.
/balka:deployStage 5, Deploy — push the branch, open the pull request, let Copilot code review it, address every finding, and print how this change lands. Merging stays with the code owner.
/balka:initSet this repo up for balka — current documentation, entry instructions, facts, proposals, opt-in hooks and review policy; optionally the deploy gate and Stage 6 detector.
/balka:intentStage 1, Plan — write intent.md for a change: the problem, the outcome and the behaviours in plain words; split a programme into children.
/balka:planStage 3, Build, first half — plan the implementation from an accepted spec.md in plan mode, naming the files that change and the scenarios the work proves.
/balka:reflectReflect — read the proposals inbox, the change directories and the git history, and produce the change list for the next version of the loop, each item routed to the plugin or kept in this project.
/balka:specStage 2, Design — turn an accepted intent.md into spec.md, the Gherkin .feature files that are the contract, and their bindings written red.
/balka:testStage 4, Test — run the verify commands once more in a fresh context, then a verifier that did not write the code checks spec, scenarios, bindings and code against each other. Produces a fix or a stop, never a file.
/balka:verifyFill the active Balka config's verify block — the build, test, lint and scenario commands Build runs until green — by proposing each from the repo and running it once.
/balka:watchStage 6 — scaffold the CI failure-rate detector into this repo: bands.yaml, the watcher script and a scheduled workflow.
bash "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/protect-scenarios.sh" "${CLAUDE_PROJECT_DIR}"
bash "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/scenario-commit.sh" "${CLAUDE_PROJECT_DIR}"
bash "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/plan-sync.sh" "${CLAUDE_PROJECT_DIR}"
bash "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/deploy-gate.sh" "${CLAUDE_PROJECT_DIR}"
Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Harness-native ECC operator layer - 71 agents, 314 skills, 98 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
200 skills · 98 commands · 71 agents · 1 MCP server · hooks
Automated code review for pull requests using multiple specialized agents with confidence-based scoring to filter false positives
1 command
Headroom startup hooks for Claude Code and GitHub Copilot CLI.
hooks
Use Alibaba Open Code Review from local Codex.
2 skills
Investigate incidents, review code and execute UAT for release readiness, scan code for vulnerabilities, and run penetration tests with AWS DevOps Agent and AWS Security Agent.
13 skills · 9 commands · 1 MCP server
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.