Embeddable Trust Badges
Display your agent’s Mnemom Trust Rating anywhere. Dynamic SVG badges that update automatically, suitable for GitHub READMEs, websites, documentation, A2A Agent Cards, and package registries.
┌─────────────────────────────────────────┐
│ score: [ Mnemom Trust | 782 ] │
│ score_tier: [ Mnemom Trust | 782 Established ] │
│ score_trend: [ Mnemom Trust | 782 ↑ ] │
│ compact: [ 782 ] │
└─────────────────────────────────────────┘
Badges are served as SVGs from a public endpoint. No authentication required. Cached for 60 minutes.
Badge Endpoint
GET https://api.mnemom.ai/v1/reputation/{agent_id}/badge.svg?variant={variant}
| Parameter | Type | Required | Description |
|---|
agent_id | path | Yes | The agent identifier |
variant | query | No | Badge variant (default: score) |
Response: image/svg+xml with Cache-Control: public, max-age=3600 (60-minute cache).
Badge Variants
1. score (default)
Displays the Mnemom label and numeric score.
Best for: GitHub READMEs, documentation sites, anywhere you want a clean trust signal.
https://api.mnemom.ai/v1/reputation/agent-xyz/badge.svg?variant=score
2. score_tier
Displays the score with the human-readable tier label.
[ Mnemom Trust | 782 Established ]
Best for: Marketing pages, product listings, anywhere non-technical audiences will see it.
https://api.mnemom.ai/v1/reputation/agent-xyz/badge.svg?variant=score_tier
3. score_trend
Displays the score with a directional trend arrow showing 30-day movement.
Best for: Dashboards, monitoring pages, anywhere trajectory matters.
https://api.mnemom.ai/v1/reputation/agent-xyz/badge.svg?variant=score_trend
Trend arrows:
- ↑ (up arrow): Score improved over the past 30 days
- ↓ (down arrow): Score declined over the past 30 days
- → (right arrow): Score unchanged
4. compact
Displays only the numeric score, no labels.
Best for: Tight layouts, inline badges, status bars, bot profile pictures.
https://api.mnemom.ai/v1/reputation/agent-xyz/badge.svg?variant=compact
Pre-Eligible Agents
Agents that have not yet reached the 50-checkpoint minimum display a progress badge instead of a score:
[ Mnemom | Building... 23/50 ]
This badge automatically transitions to the scored variant once the agent becomes eligible.
Not Rated Agents
Agents with no reputation record at all (never registered with Mnemom) display a neutral “Not Rated” badge instead of a broken image:
[ Mnemom Trust | Not Rated ]
This ensures badge embeds degrade gracefully. The badge links to the Mnemom claim page so the agent owner can register.
https://api.mnemom.ai/v1/reputation/unknown-agent/badge.svg
The “Not Rated” badge is returned with the same Cache-Control: public, max-age=3600 header as scored badges. Once the agent is registered and earns a score, the badge automatically updates.
Embed Code
GitHub README (Markdown)
The most common use case. Add a clickable badge that links to the agent’s reputation page.
[](https://www.mnemom.ai/agents/agent-xyz/reputation)
With tier label:
[](https://www.mnemom.ai/agents/agent-xyz/reputation)
HTML Websites
Standard HTML embed with link to the full reputation page.
<a href="https://www.mnemom.ai/agents/agent-xyz/reputation"
target="_blank"
rel="noopener">
<img
src="https://api.mnemom.ai/v1/reputation/agent-xyz/badge.svg?variant=score"
alt="Mnemom Trust Rating"
/>
</a>
React / JSX
function TrustBadge({ agentId }) {
const badgeUrl = `https://api.mnemom.ai/v1/reputation/${agentId}/badge.svg?variant=score`;
const pageUrl = `https://www.mnemom.ai/agents/${agentId}/reputation`;
return (
<a href={pageUrl} target="_blank" rel="noopener noreferrer">
<img src={badgeUrl} alt="Mnemom Trust Rating" />
</a>
);
}
A2A Agent Cards
Add a trust block to your A2A Agent Card for programmatic trust discovery:
{
"id": "agent-xyz",
"name": "My Agent",
"skills": ["..."],
"trust": {
"provider": "mnemom",
"score": 782,
"grade": "A",
"verified_url": "https://api.mnemom.ai/v1/reputation/agent-xyz",
"badge_url": "https://api.mnemom.ai/v1/reputation/agent-xyz/badge.svg"
}
}
Other agents can programmatically read the trust block to make delegation decisions. See Reputation in A2A Agent Cards for the full integration guide.
The score and grade in the A2A trust block are static snapshots that may become stale. Consumers SHOULD verify by fetching the verified_url for the latest score. The badge_url always returns the current score.
npm / PyPI Package READMEs
For agents distributed as packages, add the badge to your package README:
npm (package.json README):
## Trust Rating
This agent is verified by [Mnemom](https://mnemom.ai):
[](https://www.mnemom.ai/agents/agent-xyz/reputation)
PyPI (setup.cfg / pyproject.toml README):
## Trust Rating
[](https://www.mnemom.ai/agents/agent-xyz/reputation)
Independently verified Mnemom Trust Rating. [Learn more](https://docs.mnemom.ai/concepts/reputation-scores).
Documentation Sites
For Mintlify, Docusaurus, GitBook, or other documentation platforms:
## Trust Verification
This agent maintains a public Mnemom Trust Rating:
[](https://www.mnemom.ai/agents/agent-xyz/reputation)
The score is computed from independently verified integrity checkpoints.
See the [scoring methodology](https://docs.mnemom.ai/protocols/aap/reputation-methodology) for details.
Slack / Teams Bot Profiles
For conversational agents deployed in Slack or Microsoft Teams:
- Use the
compact variant for profile images or status indicators
- Include the full reputation URL in the bot’s profile description
- Reference the score in bot introduction messages:
Hi, I'm AgentBot. My Mnemom Trust Rating is 782/1000 (A - Reliable).
Verify: https://www.mnemom.ai/agents/agent-xyz/reputation
Caching Behavior
Badges are cached at multiple levels:
| Layer | TTL | Description |
|---|
| CDN edge cache | 60 minutes | Cache-Control: public, max-age=3600 |
| API cache | 5 minutes | Score recomputation cache |
| Browser cache | 60 minutes | Standard HTTP caching |
Implication: After a score change, badges may take up to 60 minutes to reflect the new value. This is intentional — frequent badge updates would increase CDN costs without meaningful benefit (scores change incrementally).
For real-time score data, use the API endpoint directly:
curl https://api.mnemom.ai/v1/reputation/agent-xyz
Dark and Light Mode
Badge SVGs use neutral colors that work on both light and dark backgrounds:
- Score label: Dark gray text on light gray background
- Score value: White text on color-coded background (green for A+, amber for B+, red for C)
- Border: Subtle border for contrast on both backgrounds
The color coding follows the grade scale:
| Grade Range | Badge Color |
|---|
| AAA, AA, A | Green |
| BBB, BB | Amber |
| B | Orange |
| CCC | Red |
| NR | Gray |
If the default badge styling does not work on your background, you can use the API response to render a custom badge. Fetch GET /v1/reputation/{agent_id} and use the grade, score, and tier fields to render your own component.
Team Badges
Teams have their own badge endpoint that works identically to agent badges:
GET https://api.mnemom.ai/v1/teams/{team_id}/badge.svg?variant={variant}
| Parameter | Type | Required | Description |
|---|
team_id | path | Yes | The team identifier |
variant | query | No | Badge variant (default: grade) |
style | query | No | light (default) or dark |
Available variants: score, grade, score_grade, score_trend, score_tier, compact
| Variant | Display |
|---|
score | [ Team Trust | 812 ] |
grade | [ Team Trust | AA ] |
score_grade | [ Team Trust | AA 812 ] |
score_trend | [ Team Trust | AA 812 ↑ ] |
score_tier | [ Mnemom Team | 812 Established ] |
compact | [ AA ] |
Pre-eligible teams (fewer than 10 assessments) show a progress badge instead of a score:
[ Team Trust | Building 4/10 ]
Teams need only 10 team risk assessments for eligibility (vs. 50 integrity checkpoints for individual agents).
Markdown:
[](https://www.mnemom.ai/teams/team-abc123/reputation)
HTML:
<a href="https://www.mnemom.ai/teams/team-abc123/reputation">
<img src="https://api.mnemom.ai/v1/teams/team-abc123/badge.svg?variant=score"
alt="Team Trust Rating" />
</a>
React:
function TeamBadge({ teamId }) {
const badgeUrl = `https://api.mnemom.ai/v1/teams/${teamId}/badge.svg?variant=score`;
const pageUrl = `https://www.mnemom.ai/teams/${teamId}/reputation`;
return (
<a href={pageUrl} target="_blank" rel="noopener noreferrer">
<img src={badgeUrl} alt="Team Trust Rating" />
</a>
);
}
Team badges use the same 60-minute CDN caching behavior as agent badges.
GitHub Action: CI/CD Reputation Gates
Use the mnemom/reputation-check GitHub Action to gate deployments on minimum reputation scores. If an agent’s score drops below the configured threshold, the workflow step fails.
# .github/workflows/deploy.yml
name: Deploy with Reputation Check
on:
push:
branches: [main]
jobs:
deploy:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Check agent reputation
uses: mnemom/reputation-check@v1
with:
agent-id: 'my-agent-id'
min-score: 700 # Minimum score (default: 600)
min-grade: 'A' # Minimum grade (optional, alternative to min-score)
fail-on-not-rated: true # Fail if agent has no score yet (default: false)
- name: Deploy
run: npm run deploy
Action inputs:
| Input | Required | Default | Description |
|---|
agent-id | Yes | — | Agent identifier to check |
min-score | No | 600 | Minimum acceptable score (0 — 1000) |
min-grade | No | — | Minimum acceptable grade (e.g., A, BBB). Overrides min-score if set. |
fail-on-not-rated | No | false | Whether to fail if the agent has no published score |
Action outputs:
| Output | Description |
|---|
score | Current reputation score |
grade | Current letter grade |
tier | Current tier label |
passed | Whether the check passed (true/false) |
Combine the reputation check with deployment protection rules to prevent deploying agents that have dropped below your organization’s trust threshold.
One-Click Copy from Dashboard
The easiest way to get embed code is from your agent’s reputation page:
- Navigate to your agent’s dashboard
- Click the reputation score card
- Click “Embed Badge”
- Select a variant and format
- Click copy — the embed code is on your clipboard
The embed modal supports all four variants and four formats (Markdown, HTML, React, A2A).
See Also