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: avoid granting destructive tools to non-admin roles (#131)
aff9639 Rudy Mizrahi Celekli committed 1 month ago
fix(cli): close connection and discovery probes before their loops exit (#153)
e9788b7 Rudy Mizrahi Celekli committed 1 month ago
fix(templates): ignore escaped braces when collecting fixture keys (#151)
cac69fb Rudy Mizrahi Celekli committed 1 month ago
fix(concurrency): recheck throttling after queued permit admission (#150)
bcf24f3 Rudy Mizrahi Celekli committed 1 month ago
fix(http): reject truncated judge replies without dropping SUT text (#149)
dc4d745 Rudy Mizrahi Celekli committed 1 month ago
fix: use private staging files for checkpoint saves (#148)
7439900 Rudy Mizrahi Celekli committed 1 month ago
fix: reject incomplete or extra ground-truth verdicts (#138)
23a67ff Rudy Mizrahi Celekli committed 1 month ago
fix: report malformed HTTP success payloads as provider errors (#125)
5b9b127 Rudy Mizrahi Celekli committed 1 month ago
fix: prevent truthiness from inflating atomic grounding scores (#137)
f9938b4 Rudy Mizrahi Celekli committed 1 month ago
fix: reject partial and unscored results in CI example (#136)
ffda13c Rudy Mizrahi Celekli committed 1 month ago
Report malformed fixture files as validation errors (#140)
bdfb8cf Rudy Mizrahi Celekli committed 1 month ago
fix: reject invalid inspection folder names during validation (#139)
fa55686 Rudy Mizrahi Celekli committed 1 month ago
fix: compare missing and unscored inspections without false gaps (#132)
4cebf80 Rudy Mizrahi Celekli committed 1 month ago
fix: reject malformed governance fields in legacy fixtures (#141)
ed79602 Rudy Mizrahi Celekli committed 1 month ago
fix: enforce judge call budget under concurrent evaluations (#135)
51da1d6 Rudy Mizrahi Celekli committed 1 month ago
fix: show unscored artifact results without false grades or verdicts (#134)
dca29c3 Rudy Mizrahi Celekli committed 1 month ago
fix: fail inspections whose consistency cap drops below threshold (#133)
5815f47 Rudy Mizrahi Celekli committed 1 month ago
fix: reject inconsistent checkpoint entries during resume (#130)
0dd5005 Rudy Mizrahi Celekli committed 1 month ago
fix: generate authorization cases using provider role IDs (#129)
0f6587a Rudy Mizrahi Celekli committed 1 month ago
fix: preserve audit inputs when saving generated fixtures (#128)
577cf14 Rudy Mizrahi Celekli committed 1 month ago
fix: stabilize generated authorization case IDs across runs (#127)
14a1421 Rudy Mizrahi Celekli committed 1 month ago
fix: compare scorecards with unscored overall results (#126)
2070eb2 Rudy Mizrahi Celekli committed 1 month ago
fix: prevent stale HTTP sources from contaminating retrieval evidence (#124)
87f19cc Rudy Mizrahi Celekli committed 1 month ago
feat: add Requesty as an OpenAI-compatible provider (#112)
82a268c Thibault Jaigu committed 1 month ago