Engineering workflow Bundle · 22 skills
- Use it when
- you want the agent to agree with you on what to build before it codes, then build it in small tested steps
- It does
- interviews you about a change, writes a spec and tickets on your issue tracker, builds each ticket test-first, then reviews the diff
- You get
- a project glossary and decision records, a spec and ordered tickets, tested commits, a review against your standards and the spec, and ideas for next time
- Needs
- a code repo, and one run of the setup skill there to record your issue tracker: GitHub, GitLab, local Markdown files or another
- Runs
- each main step only when you type it; supporting skills, such as the test-first and review ones, the agent can also start by itself
- Unlike
- frameworks that own the whole process, each step is a small skill you call yourself and can edit
- Also covers
- triaging incoming issues, debugging hard bugs, planning work too big for one session, and finding refactors that simplify a codebase
- Start with
- ask-matt, a skill that maps the others and points you to the one that fits
TLDR
Matt Pocock, a TypeScript teacher, built these skills to keep coding agents aligned with what you want and their code small and tested. A one-time setup skill records which issue tracker the repo uses and what its triage labels are called. The flow then starts with an interview: the agent asks about the change in numbered rounds, each question with its recommended answer, until no decision is left open, and writes the project's terms to a glossary and hard-to-reverse choices to architecture decision records as you go. Questions that talk can't answer go to a throwaway prototype or a background research agent. Next, the conversation becomes a spec on the tracker, then tickets: thin slices that each work end to end and list the tickets that must land first. The build step takes one ticket at a time, test-first, and ends with a review that checks the diff against your coding standards and, separately, against the spec; another skill builds a whole spec at once with parallel subagents. Finally, a retrospective proposes new checks, coding standards or agent instructions so the next run goes better.
Summarizes the repo at 5beb35b. Its skills have changed by 144 words since.
Any agent, with the skills CLI
npx skills add https://github.com/mattpocock/skills --skill setup-matt-pocock-skills --skill ask-matt --skill grill-with-docs --skill grill-me --skill grilling --skill domain-modeling --skill research --skill handoff --skill prototype --skill wayfinder --skill to-spec --skill to-tickets --skill triage --skill implement --skill implement-spec --skill tdd --skill codebase-design --skill code-review --skill diagnosing-bugs --skill retro --skill writing-for-agents --skill improve-codebase-architectureClaude Code plugin, with 5 more of the repo's skills
claude plugin install mattpocock-skills@claude-plugins-officialCodex plugin, with 5 more of the repo's skills
codex plugin marketplace add mattpocock/skills && codex plugin add mattpocock-skills@mattpocockGitHub Copilot CLI plugin, with 5 more of the repo's skills
copilot plugin marketplace add mattpocock/skills && copilot plugin install mattpocock-skills@mattpocockGemini CLI, with 5 more of the repo's skills
gemini skills install https://github.com/mattpocock/skills.git --path skills/engineering && gemini skills install https://github.com/mattpocock/skills.git --path skills/productivitySource on GitHub · All skills from mattpocock/skills
- Installs
- 1,189,452 of grill-me, its most installed skill, on skills.sh, Sep 21, 2026
- Repo stars
- 266,454 mattpocock/skills
- First seen
- Feb 10, 2026 on skills.sh, its first skill
Its skills, in the order they're used
The ones marked “Only in the bundle” need the rest of it, so they're only listed here. The others also have their own entry in the skills list.
-
Sets up a repo for the Matt Pocock skills: issue tracker, triage labels and domain docs.
Writes the tracker, label and domain-doc settings under docs/agents/ that the other skills read; on its own, nothing uses them.
-
Picks the right skill or flow from the Matt Pocock bundle for your situation.
A map of the other skills that reads each one's SKILL.md before recommending it, so it only helps once they're installed.
-
Runs the same interview and writes ADRs and a glossary as it goes.
Its whole SKILL.md loads grilling and domain-modeling.
-
Interviews you relentlessly to sharpen a plan or design.
Its whole SKILL.md loads grilling.
-
grilling
744K installsStress-tests a plan, decision or idea, one round of questions at a time.
-
domain-modeling
672K installsBuilds and sharpens the project's domain model, CONTEXT.md and ADRs.
-
research
538K installsInvestigates a question against primary sources and saves the findings as Markdown.
-
handoff
842K installsCompacts the conversation into a handoff document for another agent.
-
prototype
815K installsBuilds a throwaway prototype to answer a design question.
-
Plans work too big for one session as a map of decision tickets.
Resolves its decision tickets by loading grilling, domain-modeling, research and prototype, on the tracker that setup-matt-pocock-skills configures (local Markdown files without it).
-
Turns the conversation into a spec on the issue tracker.
Publishes to the issue tracker, with the ready-for-agent label, that setup-matt-pocock-skills configures.
-
Breaks a plan into tracer-bullet tickets with blocking edges.
Publishes to the issue tracker, with the ready-for-agent label, that setup-matt-pocock-skills configures.
-
Moves issues and external PRs through a triage state machine and writes agent-ready briefs.
Needs the tracker and label mapping from setup-matt-pocock-skills, and loads grilling and domain-modeling to flesh out a request.
-
Implements a piece of work from a spec or tickets.
Builds through tdd and closes with code-review, loading both as steps.
-
Implements a spec through its tickets, with subagents.
Works a spec and its tickets from to-spec and to-tickets on the configured tracker, has each subagent load tdd, and ends with code-review.
-
tdd
939K installsGuides test-first feature and bug work with a red-green-refactor loop.
-
codebase-design
651K installsGives a shared vocabulary and principles for designing deep modules.
-
code-review
590K installsReviews changes since a fixed point against the repo's standards and the spec.
-
diagnosing-bugs
639K installsRuns a diagnosis loop for hard bugs and performance regressions.
-
retro
116K installsRuns a retrospective on a coding session and improves the agent's setup.
-
writing-for-agents
294K installsGuides writing documents for agents: skills, AGENTS.md and CLAUDE.md.
-
Scans a codebase for deepening opportunities, renders an HTML report, then grills through the one you pick.
Loads codebase-design for its vocabulary, then grilling and domain-modeling to work through the refactor you pick.