logic-health

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a git command to identify recently changed files for auditing. This allows the agent to focus its analysis on high-impact areas of the codebase.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes untrusted data from the codebase.
  • Ingestion points: Source code files, .logic-lens.yaml configuration file, and git log command output.
  • Boundary markers: The instructions do not define specific delimiters or "ignore" instructions for the content of scanned files.
  • Capability inventory: Read access to codebase files, execution of git commands, and generation of logic health reports.
  • Sanitization: Absent; the skill does not define explicit methods for filtering or sanitizing file content.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 06:47 AM
Security Audit — agent-trust-hub — logic-health