Security context

What an agent needs to avoid regressing past fixes and find the next vuln in this repo.

53ai/53aihub

              main @ ab7c91e
            

Fixes

CVEs

MEDIUM

Peak severity

Highlights

Indirect Prompt Injection: 1 prior fix. Scrutinize any change in this area.

api/controller/relay.go: most-fixed (1 issue). Treat as high-risk during review.

0 high-severity fixes in this history; regressions here are high-impact.

Recurring patterns

The bug types that recur here, drawn from past fixes, not open vulnerabilities.

Indirect Prompt Injection: Untrusted LLM outputs (including answers and reasoning steps) could contain injected prompts that contaminate shared storage or cross-request state if persisted without redaction. This may lead to secondary injection exploits or unauthorized state exposure.