Security context

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

agentuniverse-ai/agentuniverse

              master @ 254ecd2
            

Fixes

CVEs

CRITICAL

Peak severity

33.3%

Coverage

Highlights

Denial of Service: 3 prior fixes. Scrutinize any change in this area.

agentuniverse/agent/action/knowledge/reader/cloud/confluence_reader.py: most-fixed (1 issue). Treat as high-risk during review.

8 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.

Command Injection: Arbitrary shell execution with shell=True is gated only by an opt-in flag; default False prevents accidental RCE, but any workflow enabling it without strict input validation is exposed to prompt injection.

Cypher Injection: F-string interpolation of node IDs into Cypher queries allowed injection; parameterization and strict integer-list validation are required to prevent graph database compromise.

Information Disclosure: Credentials leak into document metadata via metadata.update; sensitive key filtering is essential for all cloud readers to prevent token exposure.