Documentation
AI

Skills

How this project uses shared AI context and how to add new context files.

Existing AI context (.agents/skills)

This project keeps AI-specific guidance in .agents/skills.

  • AGENTS.md is the single source of truth.
  • It contains short global instructions and a context map.
  • Topic files in .agents/skills/*/SKILL.md hold focused context (for example auth, db, seo, i18n).

Available skills

PromptSkillShort description
/admin-tasks Add admin section, update admin tablesadmin-tasksCreating CRUD features and admin UI sections
/ai-sdk Integrate AI generation, call OpenAI modelsai-sdkUse Vercel AI SDK, generateText, streamText, tool calling
/analytics Add analytics events, configure trackinganalyticsGoogle Analytics, Plausible, PostHog, user action tracking
/api-requests Fetch data from API, debug endpointapi-requestsFetch data, debug tRPC/oRPC endpoints
/authentication Implement login, fix auth bugauthenticationLogin flows, OAuth providers, session management
/components Create UI component, update Tailwind stylescomponentsBuild reusable React components with Tailwind
/cron-jobs Add background task, configure workercron-jobsSchedule and manage background cron jobs
/dashboards Create analytics dashboard, update chartsdashboardsBuild data dashboards and chart visualizations
/data-tables Create data table, add paginationdata-tablesSortable, filterable data tables with pagination
/dates Format dates, handle timezonesdatesDate formatting, timezones, date-fns utilities
/database-operations Create Prisma migration, update schemadatabase-operationsPrisma migrations, schema changes, DB queries
/emails Send email notification, create templateemailsTransactional email templates and delivery
/environment-variables Add or update environment variablesenvironment-variablesManage .env files and runtime config
/imports-exports Export data to CSV, import bulk dataimports-exportsCSV/bulk data import and export flows
/features Scaffold new feature, add domain logicfeaturesEnd-to-end feature scaffolding and domain setup
/filters Add search filters, implement query statesfiltersURL-based filters and query state management
/formatting Update code style rules, explain formattingformattingBiome config, linting, and code style rules
/forms Create form, add validationformsReact Hook Form, Zod validation, form patterns
/infra Managing infrastructure, terraforminfraTerraform provisioning and server infrastructure
/internationalization Add translations, handle localizationinternationalizationi18n setup, translation keys, locale routing
/notifications Send in-app notifications, handle alertsnotificationsIn-app notification delivery and alert handling
/payments Add Stripe checkout, fix billing logicpaymentsStripe integration, subscriptions, billing flows
/permissions Update user roles, check permissionspermissionsRole-based access control and permission checks
/project-structure Understand app architecture, find filesproject-structureNavigate and understand the codebase layout
/database-seeds Seed database, create test datadatabase-seedsPrisma seed scripts and test data generation
/seo Update meta tags, add opengraph imagesseoSEO metadata, sitemap, OpenGraph configuration
/storage Upload files, handle S3/R2 storagestorageFile uploads with S3-compatible object storage
/tasks Manage tasks, create plan, executetasksAI task planning and step-by-step execution
/theming Update Tailwind styles, customize fontsthemingGlobal CSS variables, Tailwind theme, shadcn/ui

More skills available

You can install additional pre-built skills from skills.sh.

How to add new context

Create a file

Create a focused context file in .agents/skills/*/SKILL.md.

Keep it focused

Keep it short and specific to one domain (for example: billing, notifications, search).

Add a link to the new file in AGENTS.md under the context map.

Avoid duplication

Do not duplicate the same rules in tool-specific root files.

We follow the AgentSkills Standard for structuring our AI context files. Use a simple structure:

  • Purpose (what this context is for)
  • Key conventions
  • Important files and folders
  • Common pitfalls

This keeps token usage low and helps AI tools load only relevant context.

On this page