GitHub Action that lints, trust-scores, and estimates token savings for agent skill files on every pull request.
- Create an
mb_live_API key at modelbound.co/settings/api-keys. - Add it as a repository secret:
MODELBOUND_API_KEY. - Add
.github/workflows/modelbound-skills.yml:
name: ModelBound Skill Check
on:
pull_request:
push:
branches: [main]
permissions:
contents: read
pull-requests: write
jobs:
skill-check:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
with:
fetch-depth: 0
- uses: ModelBound/skill-check-action@v1.0.5
with:
mode: full
scan-all-on-main: true
env:
MODELBOUND_API_KEY: ${{ secrets.MODELBOUND_API_KEY }}| Input | Default | Description |
|---|---|---|
mode |
full |
lint, trust, optimize, or full |
skills-glob |
see action.yml | Globs for skill files |
publish-report |
true |
Publish public badge data to modelbound.co |
api-url |
https://modelbound.co |
API base URL |
min-trust |
0 |
Fail if average trust is below this score |
comment |
true |
Post a PR summary comment |
mcp-version |
0.4.6 |
Pin for local lint via modelbound-mcp |
scan-all-on-main |
false |
On main pushes with no changed skill files, scan all tracked files matching skills-glob (keeps README badges fresh) |
- lint — local SKILL.md lint only (no API key)
- trust — lint + cloud trust scoring
- optimize — lint + optimize dry-run estimate
- full — lint + trust + optimize + public report for README badges
After a successful full run:
[](https://modelbound.co/connect/github-actions?repo=OWNER/REPO)
[](https://modelbound.co/connect/github-actions?repo=OWNER/REPO)
[](https://modelbound.co/connect/github-actions?repo=OWNER/REPO)