Agent
Use this agent to build AI/LLM features - Claude API integration, prompt design, tool use, RAG/embeddings, streaming, evals, cost and latency tuning. Give it the feature and the quality bar.
Agent
Use this agent to design a system or feature before building it - decomposing requirements into components, choosing data models and APIs, evaluating tech trade-offs, or reviewing an existing architecture. Returns a design document, not code.
Agent
Use this agent to automate repetitive work - scripts, scrapers, data pipelines, scheduled jobs, batch file processing, API glue between services. Give it the input, the desired output, and how often it runs.
Agent
Use this agent to implement server-side code - API routes/endpoints, business logic, background jobs, third-party integrations (Stripe, email, webhooks), auth flows. Give it a specific, bounded task with acceptance criteria. Returns working code verified by real test output.
Agent
Use this agent for the business side of a product - choosing a monetization model, pricing, MVP scope cutting, competitor scan, landing page copy, launch plan. Give it the product idea and target user.
Agent
Use this agent to review a diff, branch, or module for correctness bugs, security issues, and needless complexity before merging or shipping. Read-only - it reports findings, it does not fix them.
Agent
Use this agent to write words that sell or explain - landing pages, emails, onboarding and UI microcopy, blog posts, launch/social posts, store listings. Give it the audience, the goal, and the voice if one exists.
Agent
Use this agent to answer questions with data - defining metrics and KPIs, analytics SQL, event tracking plans, funnel/retention/cohort analysis. Give it the question to answer and where the data lives.
Agent
Use this agent for database work - designing schemas, writing migrations, Postgres/Supabase RLS policies, indexes, query optimization, seed data. Give it the entities/access patterns involved.
Agent
Use this agent to investigate a bug, failing test, or unexpected behavior and find the root cause. It returns a diagnosis with evidence; it only applies the fix if the task explicitly says to.
Agent
Use this agent for deployment and infrastructure work - Vercel/Supabase deploys, CI/CD pipelines (GitHub Actions), Dockerfiles, environment variable management, domains, monitoring setup.
Agent
Use this agent to write or update documentation - READMEs, setup guides, API references, user-facing help pages, changelogs, code comments cleanup. Give it the audience and the code/feature to document.
Agent
Use this agent for the numbers side of a business - unit economics, SaaS metrics (MRR, churn, LTV/CAC), pricing math, projections, runway, fundraising-ready financials. Give it current numbers or the assumptions to model.
Agent
Use this agent to implement UI - pages, components, forms, layouts, styling, client-side state, accessibility fixes. Give it a specific screen/component with the expected behavior and any design direction. Returns working code verified by real typecheck/build output.
Agent
Use this agent to get and keep users - SEO, content strategy, distribution channels, email lifecycle, ads basics, funnel and retention diagnosis. Give it the product, target user, and current numbers if any. Returns a prioritized, experiment-framed growth plan.
Agent
Use this agent for legal groundwork on software products - terms of service, privacy policy, GDPR/CCPA basics, cookie consent, open-source licenses, contractor agreement checklists. Returns plain-language drafts and checklists. It is not a lawyer and says when a real one is needed.
Agent
Use this agent to analyze a market - TAM/SAM/SOM sizing, competitive landscape mapping, customer segmentation, trend and why-now analysis. Give it the product idea or category and the target customer. Returns a sized, sourced analysis with one beachhead recommendation.
Agent
Use this agent to implement mobile app code - React Native/Expo screens, navigation, push notifications, camera/storage, offline handling, app store release prep. Give it a specific, bounded task with acceptance criteria. Returns working code verified by real test/build output.
Agent
Use this agent when something is slow or heavy - profiling, Core Web Vitals, bundle size, slow queries, caching, load testing. It measures first and proves improvements with before/after numbers.
Agent
Use this agent to turn ideas into build-ready specs - PRDs, user stories with acceptance criteria, scope cutting, prioritization, roadmaps. Give it the idea or feature request and who it's for.
Agent
Use this agent for behavior-preserving code changes - refactors, framework/library upgrades, migrations, dead code removal, splitting oversized modules. Give it the target and the reason; tests must hold before and after.
Agent
Use this agent for research tasks - comparing libraries, reading documentation, investigating APIs/pricing/limits, summarizing long docs or logs, broad codebase exploration. Returns conclusions with sources, not raw dumps.
Agent
Use this agent for a defensive security audit of a project or feature - auth flows, access control, RLS policies, secret handling, input validation, dependency risk. Read-only - reports findings with severity and fixes.
Agent
Use this agent to write or improve tests - unit, integration, or end-to-end - for new or existing code, to reproduce a reported bug as a failing test, or to identify untested edge cases in a module.