toolskillsagentsclaude-codeengineeringmattpocockprdtddprototypedddadrskill-authoringloopsdelegationUpdated 2026-07-07

Matt Pocock's Skills (Skills for Real Engineers)

An MIT-licensed library of small, composable, model-agnostic agent skills - "Skills for Real Engineers... not vibe coding." Install: npx skills@latest add mattpocock/skills. It is the source of Kevin's 7-phase development pipeline, his forked grill-with-docs, the local engineering-flow skills kevin-engineering-flow, setup-kevin-engineering-flow, Agent Engineering Skills, Frontend and Design Skills, Security and Review Skills, Frontend and Design Skills, Frontend and Design Skills, Agent Engineering Skills, Google Workspace Skills, Security and Review Skills, Frontend and Design Skills, Browser Testing Skills, and Agent Engineering Skills, plus Frontend and Design Skills and the long-form writing trio Writing and Content Skills, Writing and Content Skills, and Writing and Content Skills. Source: github.com/mattpocock/skills, commit 0877403d1e867fd9d574117e9b34ade404f36d2a, 2026-06-30

View original tweet
View original tweet
View original tweet
View original tweet
View original tweet
View original tweet
View original tweet

Current Source Snapshot

As of the 2026-07-06 version check, mattpocock/skills is on main at 16a2a5cd00b4416f673f4ff38c7971a04dd708e7, latest release v1.0.1 (v1.0.1, tag object 385c6009a7f7b9295b8ab9f7e3e78cbe67fa1296, peeled commit 2454c95dc305c158b21a0cdafeb728879dd0359a), MIT licensed. PR #394 head is 7a4c7561d4d8d866a9ae6e0073648ac94785ac8a and adds the Fowler smell baseline to the in-progress review skill. npx skills@latest add mattpocock/skills --list --full-depth previously reported 36 skills, and the July bookmark batch shows the next rename/design pass forming around /to-spec, /to-tickets, /wayfinder, /research, and a Martin-Fowlerized /code-review. Brin previously returned caution for the repo and skill scans under conservative tolerance, so Kevin's stack should continue with curated local imports rather than a blind full-bundle install. The 2026-06-17 v1 announcement bookmark remains the key release signal: it reports a 63% reduction in skill-description token cost, splits model-invocable from user-invocable skills, adds /codebase-design, /domain-modeling, and /grilling, and rewrites /writing-great-skills from the ground up. Source: GitHub git ls-remote, 2026-07-06; X/@mattpocockuk, 2026-06-17

Philosophy

The skills are deliberately small, easy to adapt, and composable — the opposite of process-owning frameworks (GSD, BMAD, Spec-Kit) that "take away your control and make bugs in the process hard to resolve." Hack them, make them your own. They target four recurring agent failure modes: Source: repo README, 2026-06-15

Failure mode Fix Skill(s)
Agent didn't do what I want (misalignment) A grilling session — agent interrogates you before building grill-me, grill-with-docs
Agent is too verbose (no shared language) A ubiquitous language glossary the agent decodes jargon with CONTEXT.md (built into grill-with-docs)
The code doesn't work (no feedback loops) Static types, browser access, red-green-refactor tests, debug loop tdd, diagnosing-bugs
We built a ball of mud (no design care) Care about code design every day; find deepening opportunities to-prd, zoom-out, improve-codebase-architecture

Skill inventory

Engineering (daily code work): upstream ask-matt is installed locally as kevin-engineering-flow; upstream setup-matt-pocock-skills is installed locally as setup-kevin-engineering-flow; the rest of the flow uses grill-with-docs, triage, improve-codebase-architecture, to-issues, to-prd, prototype, diagnosing-bugs, tdd, domain-modeling, codebase-design, implement, and resolving-merge-conflicts. Productivity / general: grill-me, grilling, handoff, teach, writing-great-skills. In progress: decision-mapping, loop-me, review, writing-beats, writing-fragments, writing-shape. Personal/deprecated/misc: edit-article, obsidian-vault, git-guardrails-claude-code, migrate-to-shoehorn, scaffold-exercises, setup-pre-commit, plus deprecated design-an-interface, qa, request-refactor-plan, and ubiquitous-language. Kevin's current engineering import is from commit 5d78bd0903420f97c791f834201e550c765699f8. Source: GitHub tree, 2026-06-27

/setup-kevin-engineering-flow is the one-time per-repo scaffold: it configures the issue tracker (GitHub / Linear / local files), the triage label vocabulary, and where docs are saved - config the other engineering skills consume.

/loop-me is the newest high-signal addition: it interviews the user about recurring work loops and writes workflow specs for AI delegation. Kevin now has a local executable adaptation as Frontend and Design Skills, routed through the no-one-off/workflow stack rather than treated as a generic planning skill. The reviewed bookmark had no visual artifact; its value was the resolved GitHub SKILL.md and self-thread that named /loop-me as the new in-progress skill. Source: X/@mattpocockuk, 2026-06-24; Source: GitHub SKILL.md

July 2026 v1.1 Command Surface

The July batch is now a source-level command-surface update. Pocock says Skills v1.1 is cooking with /wayfinder, /to-spec, /to-tickets, improvements to /writing-great-skills, a /research skill, a Martin-Fowlerized /code-review, external-PR support for /triage, and docs for every public skill. Kevin's local response is to expose the vocabulary in Slash Command Index while keeping one executable owner per workflow. Source: X/@mattpocockuk, 2026-07-06

View original tweet
View original tweet
View original tweet
Upstream v1.1 command Kevin route now Local action
/to-spec Agent Engineering Skills Active alias. Update descriptions and Slash Command Index, but keep one spec/PRD synthesis skill.
/to-tickets Google Workspace Skills Active alias. Keep vertical-slice issue/ticket breakdown in one skill.
/wayfinder Frontend and Design Skills, Security and Review Skills, Skill Resolver Watch-only until upstream docs/source clarify whether it is a router, research planner, or navigation skill.
/research Security and Review Skills, Cloud, Data, and Service Skills, last30days research route Watch-only unless it adds a durable source-corpus, uncertainty ledger, and saved Markdown artifact contract beyond existing research flows.
/writing-great-skills Browser Testing Skills, Security and Review Skills Active route for skill authoring, no-op pruning, reference files, scripts, and progressive disclosure.
/code-review Security and Review Skills, Security and Review Skills, Security and Review Skills, Frontend and Design Skills Fold Fowler code-smell language into existing Standards/Spec review routes; do not import a duplicate reviewer by default.
/triage Browser Testing Skills Active route. The local skill already covers external PRs as dry-run first unless Kevin approves live mutation.

This command surface is generated into Slash Command Index so root agent docs can stay small. The index is the listable surface; the skill files remain the executable source.

Local command policy: expose upstream vocabulary as /COMMAND handles, but do not create a new local skill unless it owns a different repeatable workflow, has a clear source/output contract, or replaces an existing owner. /to-spec and /to-tickets are aliases; /wayfinder and /research stay watch-only; /writing-great-skills, /code-review, and /triage route to existing local owners. This keeps the command surface discoverable without turning the wiki into a drawer full of duplicate workflows. Source: User request, 2026-07-07

The local media for this cluster is mostly tweet/repo screenshot evidence rather than an executable artifact; the value is the sequence relationship across the rename posts, code-review vocabulary, and v1.1 planning notes. Source: X bookmark artifact audit, 2026-07-06

Fowler Review Baseline

PR #394 adds a curated Fowler smell baseline to the Standards axis of the in-progress review skill. The important design choice is that this is not a third review axis and not a hard lint rule. It is a Standards-axis heuristic set, with two binding constraints: documented repo standards override the baseline, and every smell is reported as a judgement call.

Kevin imported the vocabulary into Security and Review Skills rather than adding a duplicate /review command. The local review smell list is: Mysterious Name, Duplicated Code, Feature Envy, Data Clumps, Primitive Obsession, Repeated Switches, Shotgun Surgery, Divergent Change, Speculative Generality, Message Chains, Middle Man, and Refused Bequest. Source: mattpocock/skills PR #394 head 7a4c7561d4d8d866a9ae6e0073648ac94785ac8a, checked 2026-07-06

Writing-Great-Skills v1.1 Notes

The current writing-great-skills source frames skill quality around predictability, not size for its own sake. The vocabulary Kevin should keep using in Browser Testing Skills is: invocation, context load, cognitive load, router skill, information hierarchy, steps vs reference, progressive disclosure, context pointer, co-location, branch, leading word, completion criterion, legwork, post-completion steps, single source of truth, no-op, sediment, sprawl, duplication, relevance, and negation.

Local decision: /writing-great-skills continues to route to Browser Testing Skills and Security and Review Skills. Do not install a second skill-authoring command unless upstream gains a materially new executable workflow. Source: mattpocock/skills skills/productivity/writing-great-skills, commit 16a2a5cd00b4416f673f4ff38c7971a04dd708e7, checked 2026-07-06

Bundle Pattern

The repo's setup pattern is the part Kevin should copy into his own skill distribution:

  1. Install a bundle, not one giant prompt: npx skills@latest add mattpocock/skills.
  2. Force a setup skill before first use: /setup-kevin-engineering-flow records issue tracker, triage labels, and domain-doc layout.
  3. Split user-invoked from model-invoked skills. User-invoked skills orchestrate flows; model-invoked skills hold reusable discipline and can be reached automatically.
  4. Use reference files for branch depth. SKILL.md stays small; files such as DEEPENING.md, CONTEXT-FORMAT.md, tests.md, and HTML-REPORT.md load only when that branch needs them.
  5. Keep local adaptations explicit. Kevin's copies preserve upstream intent but route to local owners such as Security and Review Skills, Security and Review Skills, Security and Review Skills, Security and Review Skills, and git safety rules.

This is the minimum viable shape for Kevin's own distributed skills: bundle page, setup skill, router skill, small executable skills, branch reference files, generated registry rows, family synthesis pages, resolver rows, trigger evals, and generated packs.

The v1 split between user-invoked and model-invoked skills is the reason Kevin's local import keeps two boundaries: user-facing routers such as kevin-engineering-flow and setup flows carry Kevin-owned names, while upstream source names remain provenance. Model-invoked discipline skills can stay small and composable because they are not trying to be the command surface. Source: X/@mattpocockuk, 2026-06-17

Patterns worth adopting (the "great patterns")

  1. Ubiquitous language as a living CONTEXT.md. A glossary (not a spec): opinionated, tight one-or-two-sentence definitions, project-specific terms only, with an _Avoid_ list of rejected synonyms. Multi-context repos use a CONTEXT-MAP.md. Pocock calls it "maybe the single coolest technique in this repo" — it makes the agent more concise, names variables/files consistently, and spends fewer thinking tokens. (Mirrors Kevin's The Brain-Agent Loop compiled-knowledge move.)
  2. ADRs, offered sparingly. A 1-3 sentence record of a decision, created only when all three hold: hard to reverse, surprising without context, the result of a real trade-off. Otherwise skip it.
  3. Tracer-bullet vertical slices. to-issues and tdd both insist on thin slices that cut through every layer end-to-end (schema→API→UI→tests), never horizontal "all tests then all code." Each slice is demoable; many thin > few thick. Slices tagged HITL (needs a human) vs AFK (agent can merge solo).
  4. Deep modules + the deletion test. improve-codebase-architecture uses a strict vocabulary (module, interface, implementation, depth, seam, adapter, locality). The deletion test: imagine deleting a module — if complexity vanishes it was a pass-through; if it reappears across N callers it earned its keep. "One adapter = hypothetical seam; two adapters = real seam." (Ousterhout's A Philosophy of Software Design.)
  5. Prototype = throwaway code that answers one question. Two branches: LOGIC (a runnable terminal app to push a state machine through hard cases) or UI (several radically different variations on one route, toggled by a URL param + floating bar). Throwaway from day one; keep only the answer (into an ADR/issue/NOTES.md).
  6. PRD by synthesis, not interview. to-prd does not re-interview — it synthesizes the conversation into Problem / Solution / exhaustive User Stories / Implementation Decisions / Testing Decisions / Out of Scope, picking the highest test seam first. No file paths or code (they go stale) — except a prototype-derived snippet that encodes a decision precisely.
  7. Review along two axes in parallel sub-agents. review checks Standards (does it follow this repo's documented standards?) and Spec (does it match the originating issue/PRD?) concurrently.
  8. Skill authoring = progressive disclosure. write-a-skill: SKILL.md stays small; overflow goes to REFERENCE.md/EXAMPLES.md/scripts/. The description is the only thing the agent sees when choosing a skill, so it must encode what + when to trigger. (Same doctrine as Kevin's Browser Testing Skills.)
  9. No-op pruning is contextual deletion testing. A skill line is only valuable if deleting it changes routing, first action, tool/file choice, output contract, stop condition, or verification. Vague virtue instructions and repeated advice are skill sediment unless rewritten into checkable behavior. Source: X/@mattpocockuk thread, 2026-06-24
  10. Long-form writing splits explore from exploit. writing-fragments mines a raw pile without structure; writing-beats turns the pile into a reader journey one grounded beat at a time; writing-shape produces a coherent article paragraph by paragraph. This maps cleanly onto Kevin's content stack: gather raw material first, shape the article second, adapt for X/LinkedIn last. Source: X/@mattpocockuk, 2026-06-25; Source: GitHub in-progress skills, commit 5d78bd0903420f97c791f834201e550c765699f8

Local Adoption Matrix

Upstream skill / pattern Local route Status
grill-with-docs, grill-me, grilling Frontend and Design Skills Forked and extended with breadth-first design-tree shaping.
ask-matt kevin-engineering-flow Renamed locally so the user-invoked router belongs to Kevin's stack; upstream slug stays provenance only.
setup-matt-pocock-skills setup-kevin-engineering-flow Renamed locally as the repo-local issue/domain configuration step; adapted to AGENTS.md and Security and Review Skills.
loop-me Frontend and Design Skills Installed and updated with loop/workflow vocabulary, push-right checkpoints, briefs, and NOTES.md.
domain-modeling, deprecated ubiquitous-language Agent Engineering Skills Installed as the owner for CONTEXT.md, CONTEXT-MAP.md, and sparse ADRs.
codebase-design Frontend and Design Skills Deep-module vocabulary installed with DEEPENING.md and DESIGN-IT-TWICE.md reference files.
improve-codebase-architecture Frontend and Design Skills + Frontend and Design Skills Installed as the visual architecture-deepening report; broad plan banking still belongs to improve.
diagnosing-bugs Security and Review Skills Installed as the red-capable feedback-loop debugging route.
tdd Frontend and Design Skills Installed as red-green-refactor / vertical-slice route with tests.md, mocking.md, and refactoring.md references.
prototype Frontend and Design Skills Installed as throwaway LOGIC/UI prototype route with branch references.
to-prd / /to-spec Agent Engineering Skills Installed as conversation-to-PRD/spec synthesis route; /to-spec is the current alias.
to-issues / /to-tickets Google Workspace Skills Installed as vertical-slice issue/ticket breakdown route; /to-tickets is the current alias.
implement Security and Review Skills Installed as bounded PRD/issue execution; local version does not auto-commit unless requested.
triage, deprecated qa / request-refactor-plan Browser Testing Skills + Google Workspace Skills Installed for issue/backlog/external-PR intake; refactor plans route to issue breakdown.
resolving-merge-conflicts Agent Engineering Skills Installed with Kevin's git safety guardrails.
review Security and Review Skills, Security and Review Skills, Security and Review Skills Covered locally; preserve Standards/Spec review axes in review docs rather than adding a duplicate skill.
handoff, teach Cloud, Data, and Service Skills, Security and Review Skills Covered locally.
writing-great-skills, no-op pruning thread Browser Testing Skills, Security and Review Skills Incorporated into local skill-writing guidance; "tighten/prune this skill" now routes to skill-creator.
writing-fragments, writing-beats, writing-shape Writing and Content Skills, Writing and Content Skills, Writing and Content Skills Installed as the long-form writing pipeline: explore raw fragments, choose article beats, or shape paragraph/block drafts. Deep review on 2026-06-30 folded the upstream leading-word, grounding, pile-as-quarry, and format-debate rules into the local skill bodies.
edit-article Security and Review Skills, Cloud, Data, and Service Skills, Security and Review Skills Reference-only; local polish/publish skills already own the final editing surface.
git-guardrails-claude-code, setup-pre-commit, resolving-merge-conflicts, migrate-to-shoehorn, scaffold-exercises, obsidian-vault Existing git/process/tooling/wiki routes Reference-only unless Kevin asks for that exact workflow.

Relevance to Kevin

Kevin uses this framework as his 7 phases. The local stack now covers most of the repo through natural owners instead of a pile of overlapping commands: grill-with-docs for alignment, domain-modeling for vocabulary/ADRs, prototype for evidence, to-prd for synthesis, to-issues for tracer-bullet execution, tdd and diagnosing-bugs for feedback loops, codebase-design for deep modules, triage for intake, and loop-me for recurring work delegation.

The reviewed seven-phase visual reinforces an existing local decision rather than creating a new skill. The flow is still Grill, optional Research, optional Prototype, PRD, Issues, Implement, and Review; Kevin's only local change remains the pre-PRD breadth-first design-tree pass before high-fidelity branch work. Source: X/@mattpocockuk visual artifact, 2026-06-15

Timeline

Related Pages

The 7 Phases of AI-Powered DevelopmentDesign-Tree Exploration (Shape Before Depth)Matt PocockBrowser Testing SkillsSlash Command IndexHarness EngineeringThe Eval Loop (Slop Is an Output Problem)Skills.sh - Agent Skills Registrykevin-engineering-flowsetup-kevin-engineering-flowFrontend and Design SkillsFrontend and Design SkillsAgent Engineering SkillsFrontend and Design SkillsSecurity and Review SkillsFrontend and Design SkillsFrontend and Design SkillsAgent Engineering SkillsGoogle Workspace SkillsSecurity and Review SkillsFrontend and Design SkillsBrowser Testing SkillsAgent Engineering SkillsWriting and Content SkillsWriting and Content SkillsWriting and Content SkillsLoop LibraryNo One-Off WorkAgent Skill BundlesX Bookmark Ingest 2026-07-06

19 pages link here

Agent Skill BundlesConceptsAgent-Docs MeshConceptsBrain CapsulesMetaCodex Record & ReplayToolsDesign-Tree Exploration (Shape Before Depth)ConceptsHarness EngineeringConceptsKevin-WikiProjectsLLM Wiki Agent HarnessArchitectureMatt PocockPeopleSkill ResolverSkill ResolverSkills.sh - Agent Skills RegistryToolsSlash Command IndexMetaThe 7 Phases of AI-Powered DevelopmentWorkflowsWiki Capsules And Agent Doc PacksMetaX Bookmark Ingest 2026-06-25ResearchX Bookmark Ingest 2026-06-26ResearchX Bookmark Ingest 2026-07-06ResearchX Bookmark Ingest 2026-07-12ResearchX Bookmarks: AI Agents & Tools (Jan 2025 – Jun 2026)Tools