protocol
pass
100 / 100
Measurement report
Invokely test fixture
F
0 / 100
Scoped release decision
Requires correction within the tested scope
Evidence lanes
client
warn
96.8 / 100
security
fail
100 / 100
agent
fail
0 / 100
operations
pass
100 / 100
Agent evaluation
| Scenario | Model | Tool selection | Argument exactness | Task completion | {score} ± {margin} | Efficiency |
|---|---|---|---|---|---|---|
| echo-live-evaluation | gemini gemini-3.5-flash | 100% | 100% | 0% | 0-79% | 14824 ms |
| echo-live-evaluation | openai gpt-5.6-luna | 100% | 100% | 0% | 0-79% | 8387 ms |
| fixture-health-status | gemini gemini-3.5-flash | 100% | 100% | 0% | 0-79% | 17382 ms |
| fixture-health-status | openai gpt-5.6-luna | 100% | 100% | 0% | 0-79% | 6154 ms |
Rates include a Wilson confidence interval. A single repetition is shown as an illustrative result, not a rate.
Public reports show issue category and severity. Verified owners and Guard organizations can open finding details.
Sign inFindings
Invokely check
Dynamic input security signals were detected.
Invokely check
No dynamic output security signals were detected in this cell.
Invokely check
No dynamic output security signals were detected in this cell.