maosproject-dev/Maos AgentGate: CI Quality Gatekeeper for AI Agents
CI Gatekeeper for AI Agents. Blocks PRs if accuracy drops below threshold.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| openai_api_key | API Key for the Judge Model | yes | — |
| system_prompt | Path to your agent system prompt file | yes | prompts/system.txt |
| test_cases | Path to your golden dataset (JSON) | yes | tests/expected.json |
| threshold | Minimum passing score (0-100). Fails build if lower. | no | 90 |
| model | The LLM to test against (e.g., openai:gpt-4o-mini, openai:gpt-4o) | no | openai:gpt-4o-mini |
| temperature | LLM Temperature (0.0 = Deterministic) | no | 0.0 |
| repetitions | Number of times to run each test to smooth out noise | no | 1 |
Outputs
| name | description |
|---|---|
| score | The final evaluation score (0-100) |