Add the TaskBounty check to your AI-build workflow
If you are building with Cursor, Claude, Codex, Lovable, Bolt, Replit, or v0, the easiest habit is simple: before launch, run the local check and then ask for a human review of anything the CLI cannot verify.
Check a repo from your terminal
npx taskbounty-check@latest .Runs locally and checks GitHub Actions + CI hygiene. It writes a local report and does not upload source code or workflow contents.
Want it inside your workflow? Copy the Cursor, Claude, Codex, and GitHub Actions snippets.
Use TaskBounty in your coding agent
Run the checker as a local MCP server inside your editor, install it as an agent skill, or add it to CI. Every option is local-only: the MCP server makes no outbound network requests, never uploads source or results, and never edits files without your approval. Pin the version for reproducibility.
Cursor (.cursor/mcp.json)
{
"mcpServers": {
"taskbounty-check": {
"command": "npx",
"args": ["-y", "taskbounty-check@0.1.5", "mcp"]
}
}
}Claude Code
claude mcp add taskbounty-check -- npx -y taskbounty-check@0.1.5 mcpCodex (~/.codex/config.toml)
[mcp_servers.taskbounty-check]
command = "npx"
args = ["-y", "taskbounty-check@0.1.5", "mcp"]skills.sh (agent skill)
npx skills add eliottreich/taskbounty-checkGitHub Action (CI job summary)
permissions:
contents: read
steps:
- uses: actions/checkout@v4
- run: npx taskbounty-check@0.1.5 . --github-summary --no-network
Prompt snippets
Cursor / AI editor rule
Before I ship this app, run a local CI hygiene check:
1. In the repo root, run `npx taskbounty-check@latest .`.
2. Do not upload source code, workflow contents, secrets, or generated reports anywhere.
3. Read the local report and summarize only: candidate count, categories, and whether manual review is recommended.
4. Separate what the CLI can check (GitHub Actions + dependency update hygiene) from what still needs manual review (secrets, auth, payments, webhooks, runtime behavior).
5. Ask me before changing any files.Claude / Codex launch prompt
Act as my pre-launch safety reviewer. First run `npx taskbounty-check@latest .` locally from this repository. Treat the output as a CI/workflow hygiene signal only, not a full security audit.
Return:
- What was checked.
- What was not checked.
- The smallest safe fixes worth making before launch.
- Any items that need a human security review.
Do not transmit code or workflow contents. Do not open PRs or edit files until I approve the plan.GitHub Actions optional pre-launch workflow
name: TaskBounty local CI hygiene check
on:
workflow_dispatch:
pull_request:
paths:
- '.github/workflows/**'
- '.github/dependabot.yml'
- 'renovate.json*'
permissions:
contents: read
jobs:
taskbounty-check:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions/setup-node@v4
with:
node-version: '20'
- name: Run TaskBounty check
run: npx taskbounty-check@latest . --dry-run
README pre-launch checklist block
## Pre-launch safety check
Before sharing this app publicly, run:
```bash
npx taskbounty-check@latest .
```
This checks GitHub Actions and CI/update hygiene locally. It does not upload source code or workflow contents by default, and it is not a full app security audit. Secrets, auth, payments, webhooks, and runtime behavior still need manual review.
How to use these
- Cursor: Paste the AI editor rule into project instructions or the chat before launch work.
- Claude / Codex: Paste the launch prompt when asking the agent to review a local repo.
- GitHub Actions: Use the workflow only if you want a manual or PR-triggered CI hygiene gate. Keep the workflow least-privilege.
- README: Add the checklist block to AI-built apps where future contributors need a lightweight pre-launch habit.
The local CLI checks CI/workflow hygiene. For private repositories or broader app risks, request a confidential review. For the broader launch sequence, use the pre-launch checklist.