knowledge-citation

Pass

Audited by Gen Agent Trust Hub on May 8, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a structured workflow for generating responses with proper citations. The instructions focus on user experience and transparency without attempting to bypass security constraints.
  • [COMMAND_EXECUTION]: The skill includes a Python code block for validating file citations. This code uses the standard pathlib library to check if specific files exist on the filesystem. It does not perform any dangerous operations such as reading file contents, modifying files, or executing arbitrary shell commands.
  • [DATA_EXFILTRATION]: The skill specifies that citation metadata should be logged to a local file at /tmp/knowledge-evolution/citation-log.jsonl. This is a standard local auditing practice and does not involve any external network requests or unauthorized data transmission.
Audit Metadata
Risk Level
SAFE
Analyzed
May 8, 2026, 03:47 PM
Security Audit — agent-trust-hub — knowledge-citation