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: Surface upstream errors inside HTTP chat choices (#254)
3078878 Rudy Mizrahi Celekli committed 1 month ago
fix(gemini): keep native retries within adapter ownership (#303)
c3ff52d Rudy Mizrahi Celekli committed 1 month ago
fix: preserve exhausted-quota errors and stop MiniMax retries (#269)
d18a3b9 Rudy Mizrahi Celekli committed 1 month ago
fix: fail exact ties in full-mode analytic ensembles (#296)
89f4d5c Rudy Mizrahi Celekli committed 1 month ago
fix(gemini): fall back when models reject JSON output (#302)
e90d53c Rudy Mizrahi Celekli committed 1 month ago
fix: validate typed fixture invariants before declaring Valid (#276)
32a2cb1 Rudy Mizrahi Celekli committed 1 month ago
fix: validate audited users against all declared roles (#291)
25ddf98 Rudy Mizrahi Celekli committed 1 month ago
fix: validate audit timestamps for every B13 trail size (#295)
cedc6f6 Rudy Mizrahi Celekli committed 1 month ago
fix: preserve authoritative empty discovery permissions (#309)
b1b5c83 Rudy Mizrahi Celekli committed 1 month ago
fix: resolve hosted model creators for grader independence (#305)
c16a8ab Rudy Mizrahi Celekli committed 1 month ago
fix: distinguish inspected coverage from observed pass rate (#288)
74d4c73 Rudy Mizrahi Celekli committed 1 month ago
fix: preserve inconclusive inspection reasons in JSON reports (#294)
bd69a8f Rudy Mizrahi Celekli committed 1 month ago
fix: disclose incomplete cross-hook verification in reports (#306)
fbf8c46 Rudy Mizrahi Celekli committed 1 month ago
fix: identify invalid UTF-8 run configurations (#275)
2765a89 Rudy Mizrahi Celekli committed 1 month ago
fix(scaffold): preserve custom commands when backups conflict (#308)
84ae267 Rudy Mizrahi Celekli committed 1 month ago
fix: align authoring schemas with shipped inspection contracts (#280)
3160c51 Rudy Mizrahi Celekli committed 1 month ago
fix: isolate inspection state between overlapping API runs (#290)
171fa9a Rudy Mizrahi Celekli committed 1 month ago
fix: report atomic judge budget skips accurately (#287)
e36c485 Rudy Mizrahi Celekli committed 1 month ago
fix: reject reported LangServe judge truncation (#284)
c0122d6 Rudy Mizrahi Celekli committed 1 month ago
fix(cli): reject invalid report destinations before probing (#299)
19a2eca Rudy Mizrahi Celekli committed 1 month ago
fix(cli): reset target connection defaults when switching providers (#304)
aa51817 Rudy Mizrahi Celekli committed 1 month ago
fix(cli): enforce Full fixture policy for named aliases (#307)
cf37747 Rudy Mizrahi Celekli committed 1 month ago
fix: expose AtlasCloud in guided setup and credential detection (#258)
8066421 Rudy Mizrahi Celekli committed 1 month ago
fix: distinguish malformed MiniMax responses from empty output (#272)
dda229c Rudy Mizrahi Celekli committed 1 month ago
fix: require the AWS secret when offering Bedrock setup (#246)
c1741d7 Rudy Mizrahi Celekli committed 1 month ago
fix: exclude dotenv inline comments from credential values (#257)
b943433 Rudy Mizrahi Celekli committed 1 month ago
fix: launch setup runs through the current Python interpreter (#265)
ce5a1d0 Rudy Mizrahi Celekli committed 1 month ago
fix: Recognize and persist supported credential aliases in setup (#267)
7c36dba Rudy Mizrahi Celekli committed 1 month ago
fix: propagate failures from runs launched by setup (#273)
7925d8e Rudy Mizrahi Celekli committed 1 month ago
fix: reject empty and nonstring phrasing probes (#237)
9c3cbf3 Rudy Mizrahi Celekli committed 1 month ago