legal-intake-premature-validation-highlighting

Pass

Audited by Gen Agent Trust Hub on Apr 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides explicit instructions for the agent to execute shell commands, specifically git show, git log, and ripgrep (rg), to analyze local code repositories for bug signatures.
  • [DATA_EXFILTRATION]: The skill exposes the local directory structure and the username of a developer via hardcoded absolute file paths (/Users/adityagupta/Projects/angular-frontend). While intended for local diagnostics, these paths reveal internal environment details and make the skill non-portable.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 9, 2026, 08:57 AM