Guide · Agent skill
How to use the Convex Launch Readiness agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The convex-launch-readiness skill is an agent skill for projects built on Convex, the backend platform. It tells your AI agent to run several Convex audits, merge their findings into one report, score that report from confirmed issues, and list the fixes in priority order. This guide explains what the skill does, how to install it, how to ask for it in plain language, and what it needs before it can run.
Convex Launch Readinessopenclaw/clawhub · MIT Download the skill Quick start
- 1Download and unzip the skill. Download the ZIP from this page and unzip it so that SKILL.md sits inside a convex-launch-readiness folder.
- 2Place it in your skills folder. Move the folder into ~/.claude/skills/ for all projects, or into your project's .claude/skills/ folder for one project. You can also upload it under Settings → Capabilities → Skills in the Claude apps.
- 3Check your Convex project. Confirm the project has a convex/ directory. For the performance and error-log passes, confirm you have a deployed Convex deployment with traffic.
- 4Install the companion capabilities. Make sure convex-authz, convex-reviewer, convex-advisor, convex-insights, deploy-guard, migrate-rehearse and suggest are available, since this folder does not include them.
- 5Ask for a readiness check. Request a launch readiness check in plain language, naming Convex and the word readiness or launch.
- 6Read the report. Check the header for which passes ran and which were skipped, then review the score formula, the blockers and the ordered fix plan.
- 7Fix and re-run. Accept the fixes you want. The agent hands each one to its fixer, then re-runs the affected passes and shows the score change.
What is an agent skill?
An agent skill is a folder containing a SKILL.md file that gives an AI agent a specific set of instructions for a task. When the skill is installed and a request matches its description, the agent follows the workflow written in that file.
This skill's SKILL.md is the only file in the folder. It describes a workflow that calls other Convex audit capabilities. Those companion capabilities are not included in this download.
What the skill makes the agent do
The skill runs the Convex audits for authorization, code review, performance advice and recent error logs, then combines their results. Duplicate findings from different passes are merged, so one underlying problem is counted once in the score.
The score starts at 100. Each confirmed finding subtracts points by severity: high −15, medium −5, low −1, with a floor of 0. The report prints the formula and a per-class breakdown so the number can be reproduced. Findings the agent only suspects are listed as candidates and do not change the score.
The report groups issues as blockers (high), should-fix (medium) and nice-to-have (low). Each issue includes its evidence, where it is located and a suggested fix. The skill does not apply fixes itself. It decides what to fix and in what order, then hands off to the appropriate fixer when you ask.
Install the skill
Download the ZIP from this page and unzip it. You can install the skill in either of two places.
For all your projects, unzip it into ~/.claude/skills/. For a single project, unzip it into that project's .claude/skills/ folder. Either way, the folder should contain SKILL.md.
You can also upload the skill in the Claude apps under Settings → Capabilities → Skills.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside your project
- Claude apps: Settings → Capabilities → Skills
How to trigger the skill with natural requests
You do not need a special command. Describe what you want in plain language and mention Convex and readiness or launch. The agent matches the request to the skill's description.
Requests that should match include asking for a launch readiness check, asking for a scored report on your Convex backend, or asking which Convex problems to fix before going live.
- Run a launch readiness check on my Convex app and give me a score.
- Before we launch, what are the biggest problems in our Convex backend?
- Give me one prioritized list of Convex issues with a fix order.
- Run the Convex audits and tell me which checks you could not run.
What you need before running it
The skill needs a Convex project with a convex/ directory so the code-only audits can run. For the performance and error-log passes, it also needs access to a deployed Convex deployment that has traffic.
The skill calls companion Convex audit capabilities named convex-authz, convex-reviewer, convex-advisor, convex-insights, deploy-guard, migrate-rehearse and suggest. The source says these are not included in this folder, so you must have them available for the full run.
The skill uses the network. It runs no scripts.
- A Convex project containing a convex/ directory
- Access to a deployed Convex deployment with traffic, for the advisor and insights passes
- The companion capabilities: convex-authz, convex-reviewer, convex-advisor, convex-insights, deploy-guard, migrate-rehearse, suggest
Safety notes: scripts and network access
According to the source, this skill runs no scripts. It does use the network, because the advisor and insights passes read from a deployed Convex deployment.
The skill is designed to be read-only with respect to your code and data. It reports and dispatches; fixes happen inside the fixer capabilities, which apply their own consent and deploy-target rules. The skill's own workflow begins by classifying the target deployment as local-anonymous, dev, preview or prod, and announcing it before running passes.
A code-only run is not production-verified. The report header states which passes ran and which were skipped, and a skipped pass is recorded as a coverage gap, not a clean result.
- Runs scripts: no
- Uses network: yes
- Announces the deployment type before running checks
- Code-only results are labelled as code-only
Tips for getting useful results
Give the agent a deployed deployment with real traffic if you want the performance and error-log passes included. Without it, the report is a code-only score and says so.
After you accept fixes, ask the agent to re-run the affected passes. The skill is designed to show the score change, so you can confirm that fixes actually moved the number.
If you only want a quick look at one file, ask for a targeted review instead. The full report runs several passes and may be more than you need.
- Include a deployment with traffic for complete coverage
- Ask for a re-run after fixes to see the score delta
- Use a narrower request for a single file or a quick review
What people use it for
- Preparing a Convex app for production launch and getting one prioritized list of problems
- Aggregating results from separate Convex audit capabilities that a team already uses
- Tracking a reproducible readiness score over time, with skipped checks clearly listed
- Checking authorization and data-loss risks first, before performance and style issues
Frequently asked questions
Is the convex-launch-readiness skill free?
The source lists the licence as MIT. It does not state prices, so check the listing for any cost details. The skill also depends on companion Convex audit capabilities that the source says are not included in this folder.
How do I install the convex-launch-readiness skill?
Download the ZIP from this page and unzip it. Put the folder in ~/.claude/skills/ for all projects, or in a project's .claude/skills/ folder for one project. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
What do I need to run the Convex launch readiness check?
You need a Convex project with a convex/ directory for the code-only audits. For the advisor and insights passes, you also need access to a deployed Convex deployment with traffic. The companion audit capabilities must be available too.
Does this skill fix my Convex code automatically?
No. According to the source, the skill fixes nothing itself. It decides what to fix and in what order, then hands the work to other fixer capabilities when you accept a finding. Those fixers apply their own consent and deploy-target rules.
Why is my score only a code-only score?
The performance and error-log passes need a deployed Convex deployment with traffic. If none is available, those passes are skipped and the report says so, so the score reflects code checks only. A code-only green result means the code looks ready, not that production has been verified.
Is this skill only for Convex, or can it check other backends?
The source says it is for Convex projects only, because the audits it runs target Convex code and deployments. It is not suitable for projects built on other backends.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you are preparing a Convex app for launch and want one scored, prioritized list of problems with a clear record of which checks ran. Skip it if your project is not on Convex, if you only need a quick one-off code review, or if you expect a code-only score to prove production readiness.
Download the skill