Convex backend specialist
Writes and reviews Convex backend code (schemas, queries, mutations, actions) using Convex's current conventions.
- Repository
- openclaw/clawhub
- Stars
- 9,546
- Licence
- MIT
- Files
- 1 file
What it does
Convex is a backend platform where app logic, data storage and server functions live together. Once this skill is installed, an AI agent working on code inside a convex/ folder follows a checklist of Convex-specific rules that general AI models often get wrong, such as using object-form function definitions with argument and return validators, querying through indexes instead of full-table scans, and keeping internal functions private unless a client needs them. It also asks the agent to check that the code compiles and pushes to a Convex deployment before calling the work finished.
Great for
- Developers building an app on Convex who want fewer deploy errors from AI-written backend code
- Teams with an existing convex/ directory who want consistent schema, index and validator patterns
- Developers adding LLM chat or multi-step workflows who want the agent to use the recommended Convex components
Skip it if
- Projects that do not use Convex, since the rules are specific to its runtime and APIs and would mislead other backends
- Production databases where the agent running npx convex dev or changing deployment config could affect live data; review commands before letting it run them
- Users who are not comfortable with a command-line development setup, because the skill assumes Node tooling and a Convex login
Try asking
- “Add a new table for orders with an index on userId and a query that returns a user's recent orders”
- “Write a mutation that updates a task status and validates the status with v.union of v.literal values”
- “Set up a daily cron job that calls an action to sync data from an external API”
- “Add file upload storage to my Convex app and return a download URL from a query”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
convex-expert.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip convex-expert.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip convex-expert.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Point the agent at your existing convex/schema.ts before asking for changes so it follows your current table and index design
- Ask it to explain the commands it wants to run (npx tsc --noEmit, npx convex dev --once) and confirm which deployment they target before approving
- Use it on a development deployment first; the skill itself warns against switching a signed-in user to anonymous mode, so check your .env.local afterwards
What's inside
- SKILL.md
Redistributed unmodified from openclaw/clawhub under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.