Concepts

Badge System

Embed a live trust score badge in your README or documentation to signal your agent's security posture.

What Is the Badge?

The AgentCop badge is an SVG image served from agentcop.live/api/badge/{badge_id}. It displays your agent's current trust score and risk level, and it updates automatically when you re-scan.

Badges communicate security posture to the people who depend on your agent — teammates, security reviewers, or open-source users evaluating whether to deploy your code. A green badge is a meaningful signal. A red one is an immediate prompt to investigate.

Getting a Badge

bash
# Scan your agent and get the badge URL
agentcop scan agent.py --share

# Output:
# Trust Score: 85/100 [LOW RISK]
# Badge URL: https://agentcop.live/api/badge/abc123
# Report: https://agentcop.live/scan/abc123

Embedding in README

markdown
[![AgentCop Trust Score](https://agentcop.live/api/badge/abc123)](https://agentcop.live/scan/abc123)

Embedding in HTML

html
<a href="https://agentcop.live/scan/abc123">
  <img src="https://agentcop.live/api/badge/abc123"
       alt="AgentCop Trust Score: 85/100">
</a>

Badge Colors

Score Color Label
80–100 Green (#10b891) LOW RISK
60–79 Yellow (#f59e0b) MODERATE
40–59 Orange (#f97316) HIGH RISK
0–39 Red (#ef4444) CRITICAL

Badge via API

python
# Get badge for a scan
import httpx
badge_svg = httpx.get(f"https://agentcop.live/api/badge/{badge_id}").text
# Returns SVG string — cache with max-age: 3600

Using Badges in CI

Integrate badge generation into your CI pipeline to gate deployments on a minimum trust score.

yaml
# .github/workflows/agentcop.yml
- name: Scan with AgentCop
  run: |
    agentcop scan agent.py --min-score 70
    # Fails the build if score < 70
Note

Badges are a trust signal — not a guarantee. They show your agent passed a static scan, not that it's safe at runtime. Pair with the Runtime Monitor for complete coverage.