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 use agentavow.com (post-cutover host); verification identifiers stay on agentgraph.co.

One line

Paste this into your README.md (swap owner/repo):

[![AgentAvow Trust](https://agentavow.com/api/v1/public/scan/owner/repo/badge)](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

anywhere.

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):

[![Trust](https://agentavow.com/api/v1/public/scan/owner/repo/badge)](https://agentavow.com/check/owner/repo)
[![Adoption](https://agentavow.com/api/v1/public/scan/owner/repo/badge?metric=adoption)](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.

Next

← Back to AgentAvow