Mnemom Reputation Check

A GitHub Action that checks an AI agent's Mnemom Trust Score in your CI/CD pipeline. Gate deployments, validate agent integrity, and post trust score badges on pull requests.

Quick Start

- uses: mnemom/reputation-check@v0.4.0
  with:
    agent-id: 'your-agent-id'
    min-score: '600'

Inputs

Input	Description	Required	Default
agent-id	The agent ID to check reputation for	No*
team-id	The team ID to check team reputation for	No*
min-score	Minimum trust score required (0-1000)	No	0
min-grade	Minimum grade required (AAA, AA, A, BBB, BB, B, CCC)	No
api-url	Mnemom API base URL	No	https://api.mnemom.ai
comment	Post a PR comment with the trust score badge (true/false)	No	false

*Exactly one of agent-id or team-id is required.

Outputs

Output	Description
score	The trust score (0-1000)
grade	The trust grade (AAA-CCC or NR)
tier	The trust tier name
passed	Whether the entity passed the reputation check (true/false)
entity-type	Whether this checked an agent or team

Examples

Basic Score Check

Fail the workflow if an agent's trust score drops below 500:

name: Reputation Gate
on: [push]
jobs:
  check:
    runs-on: ubuntu-latest
    steps:
      - uses: mnemom/reputation-check@v0.4.0
        with:
          agent-id: 'agent_abc123'
          min-score: '500'

Grade-Based Gating

Require a minimum grade of A before deploying:

- uses: mnemom/reputation-check@v0.4.0
  with:
    agent-id: 'agent_abc123'
    min-grade: 'A'

PR Comment with Trust Badge

Post a trust score summary as a PR comment. Requires GITHUB_TOKEN:

name: Trust Check
on: pull_request
jobs:
  check:
    runs-on: ubuntu-latest
    permissions:
      pull-requests: write
    steps:
      - uses: mnemom/reputation-check@v0.4.0
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        with:
          agent-id: 'agent_abc123'
          min-score: '400'
          comment: 'true'

Multiple Agents

Check multiple agents in the same workflow:

name: Multi-Agent Check
on: [push]
jobs:
  check:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        agent:
          - agent_abc123
          - agent_def456
          - agent_ghi789
    steps:
      - uses: mnemom/reputation-check@v0.4.0
        with:
          agent-id: ${{ matrix.agent }}
          min-score: '600'

Use Output in Subsequent Steps

Access the trust score in later steps:

- uses: mnemom/reputation-check@v0.4.0
  id: trust
  with:
    agent-id: 'agent_abc123'

- run: echo "Agent scored ${{ steps.trust.outputs.score }} (${{ steps.trust.outputs.grade }})"

- if: steps.trust.outputs.grade == 'AAA'
  run: echo "Top-tier agent!"

Team Reputation

Check a team's collective reputation score instead of an individual agent:

Basic Team Check

- uses: mnemom/reputation-check@v0.4.0
  with:
    team-id: 'team_abc123'
    min-score: '600'

Team Grade Gate

Require a minimum team grade before deploying multi-agent systems:

name: Team Trust Gate
on: [push]
jobs:
  check:
    runs-on: ubuntu-latest
    steps:
      - uses: mnemom/reputation-check@v0.4.0
        id: team-trust
        with:
          team-id: 'team_abc123'
          min-grade: 'A'

      - if: steps.team-trust.outputs.passed == 'true'
        run: echo "Team ${{ steps.team-trust.outputs.entity-type }} scored ${{ steps.team-trust.outputs.score }}"

Team + Individual Checks

Check both the team and individual agents in the same workflow:

name: Full Trust Gate
on: [push]
jobs:
  check-team:
    runs-on: ubuntu-latest
    steps:
      - uses: mnemom/reputation-check@v0.4.0
        with:
          team-id: 'team_abc123'
          min-score: '600'

  check-agents:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        agent: [agent_1, agent_2, agent_3]
    steps:
      - uses: mnemom/reputation-check@v0.4.0
        with:
          agent-id: ${{ matrix.agent }}
          min-score: '500'

Note: Team reputation uses a 5-component model (coherence history, member quality, operational record, structural stability, assessment density). Teams need at least 10 assessments before receiving a score. See the Team Reputation Methodology for details.

Grades and Scoring

Mnemom evaluates AI agents across multiple dimensions including alignment verification, behavioral consistency, and accountability. Scores range from 0 to 1000 and map to letter grades:

Grade	Score Range	Description
AAA	900-1000	Exceptional trust
AA	800-899	Very high trust
A	700-799	High trust
BBB	600-699	Good trust
BB	500-599	Adequate trust
B	400-499	Below average trust
CCC	0-399	Low trust
NR	N/A	Not yet rated

For a detailed explanation of how scores are computed, see the Mnemom Methodology.

License

Apache 2.0 — see LICENSE.