paper-refbib-stub

Pass

Audited by Gen Agent Trust Hub on Aug 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill performs a legitimate data transformation task as described in its metadata. Behavior is consistent with the stated purpose of converting search engine results into a bibliography file.
  • [SAFE]: No network exfiltration, hardcoded credentials, or suspicious remote downloads were detected. The Python script operates entirely on local data provided via standard input.
  • [SAFE]: The script implements a _escape function using regular expressions to sanitize BibTeX special characters (e.g., {, }, \, $, &) in the input fields (title, snippet), which prevents malformed output files.
  • [SAFE]: The skill configuration uses disable-model-invocation: true, which restricts the execution environment and reduces the potential impact of processing untrusted external search data.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 7, 2026, 08:35 PM
Security Audit — agent-trust-hub — paper-refbib-stub