lightfastai

Use when a GitHub PR has CodeRabbit review-thread feedback that needs safe local triage, user-approved fixes, or a no-change review summary.

Use when a GitHub PR or branch has pending or failing PR-attached checks and needs repeated diagnosis until checks are green.

Use when a Vercel-backed release, PR merge, or promotion needs production traffic, aliases, runtime health, or related release gates verified.

Use when setting up, repairing, provisioning, or verifying local Lightfast PlanetScale database or Upstash Redis infrastructure, including db up, redis up, env files, pscale, upstash, and deleted dev-services commands.

Route protection with beforeLoad, redirect()/throw redirect(), isRedirect helper, authenticated layout routes (_authenticated), non-redirect auth (inline login), RBAC with roles and permissions, auth provider integration (Auth0, Clerk, Supabase), router context for auth state.

Framework-agnostic core concepts for TanStack Router: route trees, createRouter, createRoute, createRootRoute, createRootRouteWithContext, addChildren, Register type declaration, route matching, route sorting, file naming conventions. Entry point for all router skills.

Server-side authentication primitives for TanStack Start: session cookies (HttpOnly, Secure, SameSite, __Host- prefix), session read/issue/destroy via createServerFn and middleware, OAuth authorization-code flow with state and PKCE, password-reset enumeration defense, CSRF for non-GET RPCs, rate limiting auth endpoints, session rotation on privilege change. Pairs with router-core/auth-and-guards for the routing side.

createMiddleware, request middleware (.server only), server function middleware (.client + .server), context passing via next({ context }), sendContext for client-server transfer, global middleware via createStart in src/start.ts, middleware factories, method order enforcement, fetch override precedence.

General-purpose Linear planning and creation guidance for Lightfast MCP work. Use only when the user explicitly invokes $lightfast-linear and the task involves creating, updating, triaging, or planning Linear initiatives, projects, milestones, or issues through mcp__lightfast provider routines. Do not invoke implicitly and do not use for general Linear planning outside Lightfast MCP.

Use this skill when the user wants to write, draft, update, or revise a top-level foundation document for a company, product, or new primitive. This document captures thesis, mission, boundaries, actor model, surfaces, strategic bets, and open questions without collapsing ambiguity into implementation decisions. Applies when the user is still defining what the system is, what it is not, and what long-term direction it implies. Does NOT apply to concrete service specifications in `SPEC.md`, implementation plans, PRDs, or roadmap execution docs.

Use this skill when the user wants to write, draft, update, or revise a top-level SPEC.md service specification — the language-agnostic behavioral spec that sits at a project's root and describes what a long-running service or automation does (problem statement, goals/non-goals, components, domain model). Applies when the user says things like "write a spec for X", "draft a specification", "update SPEC.md to add Y", "revise the non-goals", "add a new component to the spec", or hands over notes and asks for them to be turned into a formal service spec. Also applies when the user references an existing SPEC.md and wants edits, even if they don't explicitly use the word "spec". Covers both greenfield creation and targeted updates of an existing file. Does NOT apply to per-topic specs under thoughts/shared/specs/ (those are produced by create_ralph_topics from a JTBD), nor to JTBDs, implementation plans, or PRDs.

Demonstrates skill triage. Select this skill for any event whose title or content mentions the word "example" explicitly.

Design, evaluate, and iterate on Claude Code agents, commands, and skills with structural best practices. Use when creating a new .claude/agents/ file, updating an existing agent's prompt, reviewing agent quality, designing a new command or skill, or when the user says "create an agent", "improve this agent", "design an agent for X", or "write a command for Y".