Independent Auditing of AI Agents. Run by human or the agent itself, to answer the most crucial question in the AI Agent Economy. Is the agent doing what is supposed to do? With iFixAi you can have this answer in less than 120 seconds.
Python agent-evaluationaiai-alignmentai-evaluationai-governanceai-safetyclidiagnostic-tooleu-ai-acthallucination-detectioniso-42001llm-evaluationllm-securitynist-ai-rmfowasp-llmprompt-injectionpythonresponsible-airisk-assessmentrisk-management
Commits on main
fix(defaults): replace shut-down Gemini and Bedrock default models (#367)
a182c0a alexcastgonz committed 20 days ago
fix: keep errored judge samples from unscoring B15/P32 FAILs (#364)
c8cb21e n-papaioannou committed 28 days ago
fix(b01): keep coverage diagnostics out of the mandatory gate's evidence floor (#353)
d3188a1 Rudy Mizrahi Celekli committed 28 days ago
fix(trajectory): report a defective delegation grant once, not once per use (#348)
079599d Rudy Mizrahi Celekli committed 28 days ago
fix(http): preserve the completed empty-output contract (#216)
6033616 Rudy Mizrahi Celekli committed 28 days ago
fix(manifest): canonicalize numeric temperatures before hashing (#323)
63b6e25 Alicia committed 28 days ago
feat(windtunnel): measure which inspections actually detect a seeded governance defect (#360)
72cbf6b Rudy Mizrahi Celekli committed 28 days ago
ci: run the test suite, and stop the documented counts from drifting (#359)
d506da2 Rudy Mizrahi Celekli committed 28 days ago
Fix P08 naive timestamp comparison TypeError — ensure _parse_timestamp returns timezone-aware UTC (#325)
3aa2af7 Ryan Vo committed 28 days ago
fix: keep malformed provider replies scoped to one probe (#247)
f361997 Rudy Mizrahi Celekli committed 28 days ago
fix(cli): report non-string config keys without a traceback (#354)
bd0ed65 Andrew Barnes committed 28 days ago
fix(artifact): show the real reason for a withheld score and only claim a cap that applied (#352)
8577585 Rudy Mizrahi Celekli committed 28 days ago
fix(scoring): keep overall scores on a grade boundary from losing the grade to float error (#351)
31b551c Rudy Mizrahi Celekli committed 28 days ago
fix(reporting): label ungraded and diagnostic evidence in the Markdown appendix (#350)
2c8801f Raphael Fakhri committed 28 days ago
fix(litellm): surface upstream mid-generation aborts instead of grading partial text (#347)
7bfe4e9 Rudy Mizrahi Celekli committed 28 days ago
fix(cli): resume full runs when ifixai.yaml saves a default suite (#346)
41f0097 Rudy Mizrahi Celekli committed 28 days ago
fix: keep errored judge samples from unscoring FAIL majorities in S02/P13/P22/P27 (#344)
510343f Rudy Mizrahi Celekli committed 28 days ago
fix(reporting): export no grade for runs without an overall score (#343)
5d8ea14 Rudy Mizrahi Celekli committed 28 days ago
fix(judge): match fallback provider keys case-insensitively (#342)
354fc88 Rudy Mizrahi Celekli committed 28 days ago
fix: compute C02/C05/C11/X04/X11 confidence intervals over the scored subset (#341)
8424323 Rudy Mizrahi Celekli committed 28 days ago