# Skill catalogue slipstream ships 78 Claude Code agent skills. Each is a `SKILL.md` with valid `name` and `description` frontmatter; each shipping skill carries a verification gate. Run `npx slipstream validate` to list them with live per-category counts. ## Shipping integrations ### frontend (11) frontend-vite-react, frontend-tailwind, frontend-router, frontend-forms, frontend-component-library, frontend-responsive-layout, frontend-dark-mode, frontend-design-system, frontend-hero-section, frontend-motion, frontend-marketing-sections ### backend (5) backend-hono-api, backend-zod-validation, backend-error-handling, backend-rate-limit, backend-openapi ### supabase (7) supabase-init, supabase-schema, supabase-rls, supabase-auth, supabase-storage, supabase-typegen, supabase-edge-function ### cloudflare (6) cloudflare-worker, cloudflare-pages, cloudflare-d1, cloudflare-kv, cloudflare-r2, cloudflare-secrets ### vercel (4) vercel-link, vercel-env, vercel-preview, vercel-deploy ### resend (4) resend-setup, resend-domain, resend-transactional, resend-webhook ### auth (4) auth-session, auth-oauth, auth-rbac, auth-password-reset ### payments (4) payments-stripe-setup, payments-checkout, payments-subscriptions, payments-webhooks ### seo (4) seo-meta-tags, seo-open-graph, seo-sitemap, seo-structured-data ### analytics (3) analytics-plausible, analytics-events, analytics-web-vitals ### git (5) git-init-repo, git-conventional-commit, git-feature-branch, git-pull-request, git-release-tag ## Discipline skills ### memory (3) memory-capture, memory-recall, memory-prune ### context (18) scoped-read, context-budget, compact-and-offload, using-slipstream, think-before-coding, systematic-debugging, brainstorm-spec, write-plan, test-driven-development, verification-before-completion, requesting-code-review, receiving-code-review, subagent-driven-development, finishing-a-branch, writing-skills, executing-plans, dispatching-parallel-agents, using-git-worktrees These are the token-discipline and methodology skills, original to slipstream. The core loop: `using-slipstream` (recall memory and read the map first, work deliberately, record what is durable), `think-before-coding` (surface assumptions and a verifiable success criterion before editing), `systematic-debugging` (a four-phase root-cause process), `brainstorm-spec` (refine a vague request into an agreed spec), `write-plan` (decompose into small verifiable tasks) and `executing-plans` (work that plan task by task with a checkpoint each). Discipline gates: `test-driven-development`, `verification-before-completion`, `requesting-code-review`, `receiving-code-review`, `finishing-a-branch`. Orchestration: `subagent-driven-development` (one task per agent, review between), `dispatching-parallel-agents` (fan genuinely-independent work out at once) and `using-git-worktrees` (isolate parallel or risky strands). `writing-skills` is the meta-skill for authoring more. ## Verification gates by example - `vercel-deploy` uses a `healthcheck` gate: it curls the production URL and requires a 200 before declaring the deploy healthy. - `supabase-typegen` uses a `command` gate that regenerates types and checks they compile. - `frontend-vite-react` uses a `build` gate. - The `memory` and `context` skills carry no gate, because they manage retrieval and memory rather than producing a deployable artifact. See [Writing a skill](Writing-a-Skill) to add your own, and [Integrations](Integrations) for how the integration skills fit a real deployment. --- SarmaLinux . sarmalinux.com . [Repository](https://github.com/sarmakska/slipstream)