Add a trust badge
AgentAvow Docs · agentavow.com
Add a trust badge to your README
Show that your tool is safe — with proof. An AgentAvow trust badge renders your repo's current signed safety score — a 0–100 number and tier word (Trusted → Blocked), coloured green→red — and clicking it opens the full, verifiable report. It's free, needs no account, and regenerates on every view, so it never goes stale.
Staged rebrand doc. Serving URLs useagentavow.com(post-cutover host); verification identifiers stay onagentgraph.co.
One line
Paste this into your README.md (swap owner/repo):
[](https://agentavow.com/check/owner/repo)
That's it. First render triggers a scan; after that it's cached and refreshed automatically.
The badge endpoint
GET https://agentavow.com/api/v1/public/scan/{owner}/{repo}/badge
- Returns an SVG (shields.io-compatible), served with
Access-Control-Allow-Origin: *so it embeds
anywhere.
- Shows the composite trust score if the repo is imported, else the security-scan score.
- Regenerates on request — it will not decay to "not scanned" in a stranger's README.
Two badges: trust and adoption
The score answers *is it safe?* Add ?metric=adoption for the second, separate signal — *do independent parties rely on it?* (downloads / dependents / stars, never a fabricated number):
[](https://agentavow.com/check/owner/repo)
[](https://agentavow.com/check/owner/repo)
The adoption badge reads "Adopted: 22.6k ★" (or "Adoption: new" when there's no signal yet). Adoption and trust are deliberately two badges — popular is not the same as safe, and one never inflates the other.
Other formats
Prefer HTML or reStructuredText? Generate any style from the badge builder at agentavow.com/badge, which outputs Markdown, HTML, and RST for four styles (compact, detailed, minimal, flat-square) and light/dark themes.
HTML
<a href="https://agentavow.com/check/owner/repo">
<img src="https://agentavow.com/api/v1/public/scan/owner/repo/badge" alt="AgentAvow Trust" />
</a>
Gate your CI on it
The badge is the display; the GitHub Action is the enforcement. Run the scan on every pull request and fail the build if the score drops below a threshold:
# .github/workflows/agentavow.yml
- uses: AgentAvow/trust-scan-action@v1
with:
repo: ${{ github.repository }}
min_score: 60 # block merges under 60/100
Or from the CLI (ships with agentgraph-sdk):
pip install agentgraph-sdk
agentgraph scan owner/repo
The viral loop
Every README reader sees the badge → clicking it lands on that repo's full report → whose primary action is minting *their own* badge. Each adoption seeds the next.