mattpocock-skills
claudeMatt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Full-stack agent framework for Claude Code, Codex, Cursor, Gemini — 39 slash commands, 46 specialist agents, 34 skills, 27 hooks. Karpathy principles (trace test, YAGNI, orphan rule, anti-stop) on every code agent. Adversarial debate triads, two-stage review (spec compliance + code quality), forensic /investigate, /elicit (18 reasoning methods), checkpoint-preview, budget circuit breaker, adversarial input hardening, continuous learning with auto-nudge, Growth OS marketing pipeline, automated PM, /audit-self regression detection.
Plugin forgebee from marketplace forgebee in forbee-dev/ForgeBee. Commands read from the repo's manifest — 2 days 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.
audit-selfbrainstormingcheckpoint-previewcode-judgecode-skepticcontinuous-learningelicitationfigma-code-syncforgebee-helpforgebee-setupgrillinvestigateproject-routerrequirements-advocaterequirements-judgerequirements-skepticreview-accessibilityreview-allreview-apireview-best-practicesreview-code-stylereview-codereview-databasereview-docsreview-performancereview-promptreview-securityreview-testsreview-wordpressstrategy-advocatestrategy-judgestrategy-skepticsurface-ambiguityterse-report/forgebee:analyticsAnalytics setup and analysis — event tracking, dashboards, funnel analysis, and data-driven insights
/forgebee:architectArchitecture advisor — design decisions, trade-offs, and technical strategy
/forgebee:audit-selfRe-run the ForgeBee self-audit — scores every skill, agent, and command against the scorecard. Detects regressions since the last run. Outputs timestamped findings.
/forgebee:auditQuery the governance audit trail — permission decisions, debate rulings, verification results, agent dispatches, and escalations.
/forgebee:browser-debugBrowser debugging specialist — Chrome DevTools, network analysis, console errors, DOM inspection, and rendering issues
/forgebee:codemapsGenerate or refresh token-lean architecture documentation optimized for AI context consumption
/forgebee:competitiveCompetitive intelligence — research competitors, build battlecards, compare positioning, and identify differentiation opportunities
/forgebee:contentQuick content production — write a single piece or small batch without the full Growth OS pipeline. Uses brand voice and hooks if available, skips strategy phases.
/forgebee:debugSystematic debugger — reproduce, isolate, diagnose, fix
/forgebee:deployDeployment coordinator — pre-flight checks, rollout, and post-deploy verification
/forgebee:design-systemDesign system operations — audit a Figma component page against code, onboard a newly built component, or audit token drift
/forgebee:docsDocumentation writer — API docs, guides, READMEs, and ADRs
/forgebee:elicitStress-test the most recent plan, design, or architecture artifact using a named reasoning method (Pre-mortem, Red Team, Inversion, Stakeholder Round Table, etc.)
/forgebee:evolveAnalyze instincts and cluster them into skills, commands, or agents
/forgebee:grillInterview the user about a plan, spec, or design in rounds, one recommended answer per question, until every decision is settled or accepted as a risk
/forgebee:growthGrowth OS orchestrator — delegates through Brand → Intel → Audience → Content Strategy → Hooks → Debate → Creation → Distribution → Measure. Never executes tasks directly; connects the dots and ships requirements to marketing specialist agents.
/forgebee:gtmGo-to-market planning — launch checklists, positioning, channel strategy, and timeline
/forgebee:ideaIdea-to-product agent with adversarial debate — validate an idea, stress-test it through Advocate/Skeptic/Judge debate, define MVP scope, plan tech stack, and create implementation roadmap
/forgebee:instinct-exportExport instincts to a shareable file
/forgebee:instinct-importImport instincts from a file
/forgebee:instinct-statusShow learned instincts with confidence scores
/forgebee:investigateForensic investigation — produces a structured case file with evidence-graded findings (Confirmed / Deduced / Hypothesized) before any code changes. Hand off to debugger-detective for the fix.
/forgebee:landingLanding page builder — conversion-optimized pages with hero, features, social proof, pricing, and CTA sections
/forgebee:launchProduct launch execution — Product Hunt, Hacker News, press outreach, and launch day coordination
/forgebee:learnAnalyze session observations and extract patterns as instincts
/forgebee:migrateMigration specialist — upgrades, version transitions, and data migrations
/forgebee:paymentsPayment integration specialist — Stripe, LemonSqueezy, Paddle setup, subscription management, webhooks, and billing logic
/forgebee:perfPerformance optimizer — profiling, bottleneck detection, and optimization
/forgebee:planBMAD-inspired planning agent — enforces a phased artifact chain from problem brief through sprint stories before any code gets written
/forgebee:pmProject management dashboard — auto-reads state.yaml and regenerates markdown views. Shows feature status, story progress, decision history, and blockers. Fully automated — syncs TASKS.md, regenerates dashboards, and surfaces what needs attention.
/forgebee:refactorCode refactoring specialist — improve structure without changing behavior
/forgebee:releaseCut a release — pre-flight gate, version bump across all manifests, changelog, tag, and PR. Wraps bump-version.sh; never reimplements version logic.
/forgebee:review/forgebee:security/forgebee:seo/forgebee:social/forgebee:team/forgebee:test/forgebee:workflowarchitectAdvises on architecture decisions, trade-offs, and technical strategy. Use for system design, technology selection, scalability planning, migration strategy, or ADRs.
audience-architectUse when defining ICP, building personas, mapping buyer journeys, running Jobs-to-be-Done analysis, or segmenting audiences.
backend-engineerBuilds APIs, server logic, middleware, auth, and business logic. Use for backend work; detects the stack from triage and delegates to wordpress-backend, n8n-builder, or llm-app-engineer, else handles directly.
brand-strategistUse when defining brand strategy, positioning, voice/tone, or messaging frameworks — archetypes, voice guidelines, messaging pillars.
content-creatorUse to write ready-to-publish content — social posts (LinkedIn, X, IG, Reels/TikTok/YouTube scripts) and long-form (landing pages, blogs, docs, READMEs, changelogs, case studies, ad copy). Routes CMS work to wordpress-content / nextjs-content.
content-strategistUse to design the content engine — pillars, topic clusters, content pyramid, platform map, angle mining, repurposing chains, series, and the editorial calendar (70/20/10). Hands briefs to content-creator.
dashboard-generatorRegenerates the markdown PM dashboards (project index, per-feature pages, decision log) from docs/pm/state.yaml. Use at the end of /workflow, /idea, /plan, or /pm runs.
database-specialistDesigns schemas, writes migrations, optimizes queries, and models data. Use for database work; detects ORM/platform from triage and delegates to supabase-specialist, else handles directly.
debugger-detectiveFinds and fixes the root cause of bugs with forensic evidence, a Failure Capture template, and the 3-failed-fix Iron Law. Use proactively when errors occur, tests fail, or bugs need reproducing.
deep-researcherResearches documentation, GitHub issues, library APIs, and technical questions with tiered, dated sources. Use when you need a verified, cited answer instead of a guess.
delivery-agentPackages finished work — consumes verification evidence, writes changelog/release notes, updates docs, checks deployment readiness. Use when /workflow reaches delivery or work needs final packaging.
devops-engineerBuilds deployment pipelines, containers, and server infrastructure — Docker, CI/CD, VPS hardening, SSL, firewalls, cloud. Use for deployment or infrastructure operations.
directOpt-out main-thread agent. Run `claude --agent forgebee:direct`, or set "agent" to "forgebee:direct" in your settings, to turn off orchestrator mode. Never dispatch it as a subagent.
email-strategistUse when designing email automation flows, segmentation, subject lines, deliverability, or lifecycle sequences (welcome, nurture, cart recovery, re-engagement, win-back).
engagement-strategistUse when growing engagement, building communities, or creating engagement playbooks — comment strategies, reciprocity loops, DM flows.
flutter-expertBuilds Flutter widgets and Dart code with Riverpod, Bloc, or Provider state management. Use for cross-platform UI on mobile, web, or desktop.
frontend-specialistBuilds UI components, styling, state management, and client-side logic. Use for frontend work; detects the framework from triage and delegates to nextjs-frontend, wordpress-frontend, flutter-expert, or ios-expert.
growth-engineerUse for growth loops, flywheels, viral mechanics, funnel diagnosis, and CRO (landing pages, forms, checkout, A/B test design). Routes WooCommerce/SaaS CRO to woocommerce-cro / saas-cro.
hook-engineerUse when creating stop-scrolling hooks, viral formulas, pattern interrupts, engagement triggers, or platform-specific hook libraries.
ios-expertBuilds native Apple apps with Swift and SwiftUI. Use for SwiftUI/UIKit, Xcode configuration, Core Data or SwiftData, CloudKit, StoreKit, or App Store submission.
llm-app-engineerBuilds LLM application code — LangChain/LangGraph, LlamaIndex, Anthropic/OpenAI SDKs, Vercel AI SDK. Use for agents, tool calling, RAG pipelines, structured output, and prompt evals in Python or TypeScript.
market-intelUse when researching competitors, building Fact-Impact-Act battlecards, comparing positioning, or analyzing niches and market trends.
marketing-analystUse to measure marketing performance — North Star/input/health metric trees, KPI dashboards, campaign analysis, attribution, A/B significance. Not code performance (that is performance-optimizer).
n8n-builderBuilds n8n workflows for integrations, automations, AI/RAG agents, and data pipelines. Use for n8n workflow JSON, webhook handling, API integrations, or low-code automation.
nextjs-contentCreates MDX content, Velite/Fumadocs/Contentlayer collections, frontmatter, and static generation in Next.js. Use when content-creator detects Next.js.
nextjs-frontendBuilds Next.js App Router pages, Server/Client Components, Server Actions, middleware, and Supabase SSR integration. Use for Next.js UI work.
nextjs-seoImplements Next.js Metadata API, sitemap.ts, robots.ts, OG image generation, JSON-LD, and next-seo (Pages Router). Use when seo-specialist detects Next.js.
orchestratorMain-thread agent that ForgeBee sets by default (plugin settings.json). Never dispatch it as a subagent. The empty body keeps Claude Code's system prompt; hooks/scripts/orchestrator-mode.js adds the rules each turn and blocks main-thread edits.
performance-optimizerProfiles and optimizes code performance — CPU, memory, queries, bundles, render. Use when something is slow or a bottleneck needs a measured before/after fix.
phpunit-engineerWrites WordPress PHPUnit tests — WP_UnitTestCase, bootstrap, factories, ACF mocking, REST and AJAX tests. Use for WordPress PHP testing; called by test-engineer when phpunit is detected.
saas-croOptimizes SaaS landing pages, pricing pages, signup flows, and A/B tests in React/Next.js. Use when growth-engineer detects a SaaS (non-WooCommerce) Node project.
scrum-masterBreaks features into context-rich stories any agent can pick up, with estimates and dependencies. Use for story decomposition, backlog grooming, estimation, or sprint planning on L/XL work.
security-auditorAudits code for security flaws across OWASP Top 10, secrets, and dependency CVEs. Use after changes to auth, data handling, APIs, or user input; delegates WordPress checks to wordpress-security.
seo-specialistRuns keyword research, on-page optimization, technical SEO audits, and pillar/cluster SEO. Routes to wordpress-seo or nextjs-seo by stack. Use for search ranking work.
session-librarianMaintains project institutional knowledge — session summaries, CLAUDE.md curation, learnings, and context recovery. Use when organizing session history or managing CLAUDE.md memory.
spec-writerWrites a build spec from a ticket or prompt — goal, outcome, verified ticket claims, approach, workstreams, and acceptance criteria with verify commands. Use in /team before any code; writes no code.
supabase-specialistBuilds Supabase backends — schemas, RLS policies, Edge Functions, Auth, Realtime, Storage, migrations. Use for Supabase work or Postgres with RLS in general.
tdd-enforcerEnforces RED-GREEN-REFACTOR — specifies tests before code, then audits order, coverage, and test quality. Use when TDD discipline is required during implementation or /workflow execution.
test-engineerWrites, fixes, and extends unit, integration, and e2e tests and closes coverage gaps. Use for test work; detects the framework from triage and delegates to phpunit-engineer, else handles directly.
ux-designerDesigns user flows, wireframes, interaction states, and accessibility requirements as UX specs. Use for flow design or UX/accessibility audits; writes no code — hands off to frontend-specialist.
verification-enforcerVerifies task completion with captured evidence — test, build, lint, API, and migration output plus exit codes. Use before marking any story or task done; code review alone does not count.
woocommerce-croUse when optimizing WooCommerce checkout, product pages, cart recovery, or e-commerce funnels. Covers WooCommerce-specific hooks and filters.
wordpress-backendUse when building WordPress plugin logic, REST endpoints, hooks, AJAX, or Settings API in PHP. ACF/SCF field architecture specialist — field groups, JSON sync, keys, meta storage, PRO-to-SCF migration.
wordpress-contentUse when creating WordPress Gutenberg block patterns, ACF-driven content, custom post type entries, WooCommerce product descriptions, or editor formatting.
wordpress-frontendUse when developing WordPress block/classic themes, template hierarchy, theme.json, or template parts.
wordpress-securityUse for WordPress security audits — sanitization/escaping, nonce verification, capability checks, SQL injection prevention, WPCS compliance.
wordpress-seoConfigures WordPress SEO — Yoast/RankMath programmatic control, XML sitemaps, permalinks, WP schema markup, WooCommerce product SEO. Use when seo-specialist detects WordPress.
wp-design-systemUse for WordPress block design systems — theme.json tokens, Figma-variables-to-presets pipeline, patterns/variations as components, Code Connect, token drift audits, theme.json vs SCSS layer ownership.
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/session-load.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/project-triage.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/load-context-rules.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/task-sync.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/pm-sync.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/learn-digest.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/load-index.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/orchestrator-mode.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/skill-activator.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/learn-recall.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/orchestrator-mode.js
node ${CLAUDE_PLUGIN_ROOT}/skills/continuous-learning/scripts/observe.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/audit-trail.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/permission-guard.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/dev-server-blocker.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/git-push-reminder.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/secret-scan.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/suggest-compact.js
node ${CLAUDE_PLUGIN_ROOT}/skills/continuous-learning/scripts/observe.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/post-edit-format.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/post-edit-typecheck.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/post-edit-console-warn.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/checkpoint.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/session-save.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/self-improve.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/console-log-audit.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/task-sync.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/pm-sync.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/audit-trail.js
node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/learn-auto.js
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
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
Build powerful, stateful CLI interfaces for any GUI application using the cli-anything harness methodology.
5 commands
Automates browser interactions for web testing, form filling, screenshots, and data extraction
12 skills
Use Alibaba Open Code Review from local Codex.
2 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.