skillZs
★ LIVE SKILL TAGS ★
>>> LIVE SKILLS INDEX <<<
* OPEN SOURCE *
NO LOGIN, NO TRACKING
※ REAL INSTALL DATA ※
← back to all skills
justbecauselabs/agent-skills99 installs

convex

Convex backend guidance for queries, mutations, actions, schema, validators, auth, components, agents, search, testing, production config, and client/server integration. Use whenever interacting with a Convex backend, including Convex functions, schema changes, HTTP endpoints, scheduling, storage, AI features, or framework integration.

How do I install this agent skill?

npx skills add https://github.com/justbecauselabs/agent-skills --skill convex
view source ↗

Is this agent skill safe to install?

  • Gen Agent Trust Hubpass

    This skill provides comprehensive documentation and coding standards for building backends on the Convex platform. It emphasizes security best practices such as server-side authentication and strict data validation.

  • Socketpass

    No alerts

  • Snykpass

    Risk: LOW · No issues

What does this agent skill do?

Convex Guidelines

This skill covers the current Convex docs surface. Start with the most relevant local guide below. If the task touches a newer platform area that is not mirrored in references/ yet, consult the linked live docs from https://docs.convex.dev/llms-full.txt before making changes.

Core rules

  1. Prefer a single exported Convex function per file unless the surrounding codebase already uses another pattern. Small private helpers in the same file are fine.
  2. Use precise validators when they improve safety, generated API types, OpenAPI output, or static codegen. Do not claim that args or returns validators are mandatory in every Convex function; when omitted, Convex falls back to v.any().
  3. Treat v.any() as an intentional escape hatch for genuinely dynamic payloads or trusted external webhook payloads, not as the default choice for normal application data.
  4. Choose the runtime deliberately. Queries and mutations stay transactional around ctx.db; actions are for external I/O and Node-only work; vector search is action-only.
  5. Prefer indexes and schema design over broad filtering. If a query shape matters for correctness or scale, verify the relevant index strategy.

Use local reference guides first

Function Guidelines

Read when: Writing or modifying Convex functions, setting up HTTP endpoints, defining validators, registering functions, designing API structure, handling pagination, or choosing between query, mutation, action, internal function, and HTTP action.

Validator Guidelines

Read when: Defining argument or return validators, especially for integers, records, unions, or dynamic payloads. Keep in mind that validators are recommended for stronger types, but not universally required by Convex.

Schema Guidelines

Read when: Creating or modifying convex/schema.ts, adding tables, defining indexes, working with _id and _creationTime, or designing table relationships.

TypeScript Guidelines

Read when: Working with Convex TypeScript types, especially Id<>, Record, discriminated unions, generated API types, static codegen tradeoffs, or Node type setup.

Query Guidelines

Read when: Writing database queries, filtering results, ordering data, using pagination, or reasoning about index-backed access patterns.

Mutation Guidelines

Read when: Updating, replacing, inserting, or deleting documents using ctx.db.

Action Guidelines

Read when: Writing actions that call external APIs, use Node modules, run vector search, or coordinate non-transactional work.

Scheduling Guidelines

Read when: Setting up cron jobs or scheduled work with scheduled functions and recurring jobs.

File Storage Guidelines

Read when: Working with uploads, downloads, generated files, serving files, or file metadata.

Full Text Search Guidelines

Read when: Implementing text search with withSearchIndex.

Migrations Guidelines

Read when: Planning or running migrations, handling schema evolution, or using this repo's migration shortcuts and run-all workflow.

Examples

Read when: You need a fuller reference implementation, including schema design, public/internal functions, and background processing patterns.

Authentication Guidelines

Read when: Configuring Convex Auth, Clerk, Auth0, WorkOS/AuthKit, custom OIDC or JWT providers, storing users in Convex, or reading auth state in functions.

Components Guidelines

Read when: Installing, using, or authoring Convex components.

AI and Agents Guidelines

Read when: Working with the Convex Agent component, threads, tools, streaming, RAG, usage tracking, workflows, the Convex MCP server, or AI codegen workflows.

Vector Search Guidelines

Read when: Defining vector indexes, generating embeddings, or running hybrid/vector retrieval.

Testing Guidelines

Read when: Adding backend tests, CI coverage, local backend tests, or convex-test mocks.

Production Guidelines

Read when: Editing convex.json, environment variables, preview deployments, multiple repositories, log streams, exception reporting, or deployment/runtime configuration.

Client Integration Guidelines

Read when: Wiring Convex into React, React Native, Next.js App Router, Next.js Pages Router, TanStack Start, Bun, Node.js, Python, Swift, Kotlin, or OpenAPI-based clients.

Practical guidance from the current docs

  1. If a task affects the public API shape, generated clients, OpenAPI output, or static codegen, strongly prefer explicit args and returns validators.
  2. If a task is internal-only and the payload is intentionally loose, a targeted v.any() can be acceptable, but make the trust boundary clear through surrounding helper names and types.
  3. If a task involves advanced search, choose between full-text search and vector search early because the schema, indexes, and function types differ.
  4. If a task involves frontend integration, pick the framework-specific client docs instead of assuming React defaults apply everywhere.
  5. If a task involves deployment or AI tooling, check convex.json and CLI docs before inventing project-level conventions; current Convex docs cover functions, aiFiles, Node runtime settings, and static codegen options there.
  6. If the local reference files are missing a detail, fall back to https://docs.convex.dev/llms-full.txt and the specific linked page for the feature you are using.

Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.

<a href="https://skillzs.dev/skills/justbecauselabs/agent-skills/convex">View convex on skillZs</a>