Skip to content
Get packs

Convex self-healing fix loop

Triage a production error, prepare a certified fix PR, and check it stops recurring, with a human merging

Source Download skill
Repository
openclaw/clawhub
Stars
9,546
Licence
MIT
Files
1 file

What it does

This skill teaches an AI agent a multi-step process for handling errors in a production Convex backend. The agent captures and triages the error, finds its root cause, writes a fix on a separate branch, and tests that fix against the backend's own checks (type checking, a rehearsal on a preview copy of the database, and replaying the failing input). It then opens a pull request for a human to review and merge, and afterwards checks that the error stops happening. It is designed never to merge or deploy to production by itself.

Great for

  • Developers running a Convex backend who want faster triage of recurring production errors
  • Teams that already use the sentinel error-capture companion skill and want a structured repair workflow
  • Engineers who want fix proposals with test evidence attached before review

Skip it if

  • Teams without a Convex deployment or without the sentinel skill installed: the SKILL.md says the loop stops if capture is absent, so it will not work as described
  • Anyone wanting unattended fixes to production: the skill explicitly keeps the human merge gate, and automated repairs to security-sensitive or destructive code are deferred to people
  • Ambiguous or hard-to-reproduce bugs: the skill instructs the agent to stop and report when root cause is unclear, so it will not produce a fix in those cases

Try asking

  • “Triage the recurring prod_error events from the last 24 hours and tell me which ones are code defects”
  • “Investigate the sentinel error for the createOrder mutation, find the root cause and prepare a certified fix PR”
  • “After my last deploy, check whether the validation error on the users table has stopped recurring”
Step-by-step guide: how to use Convex self-healing fix loopInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    convex-self-heal.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip convex-self-heal.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip convex-self-heal.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 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

  • Set the standing consent in deploy-guard first, listing which fix classes the loop may prepare automatically, since the default is limited to validator, index, ownership-check and non-destructive backfill fixes
  • Review the certification evidence in each PR (tsc result, rehearsal outcome, reproduced-then-gone check) before merging, because the skill's guarantees depend on those checks actually running
  • Keep the sentinel and preview environments in your own deployment so error data stays under your control, as the skill intends

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.