skill-optimizing

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes other instruction-based files as input, which is a standard surface for indirect prompt injection.\n
  • Ingestion points: Target skill file path and related reference files specified in the inputs.\n
  • Boundary markers: None explicitly defined within the instructions to separate input data from agent instructions.\n
  • Capability inventory: Local file reading and writing limited to modifying skill files and the project's AGENTS.md index.\n
  • Sanitization: No explicit content sanitization or validation of the ingested skill content is described in this skill document.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 11:06 AM
Security Audit — agent-trust-hub — skill-optimizing