copy-editing

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill is entirely instructional and contains no executable code or scripts.
  • [SAFE]: No network connectivity or external data transfers are initiated by the skill.
  • [SAFE]: Accessing local brand context files is a standard practice for this use case and does not involve sensitive data exposure.
  • [PROMPT_INJECTION]: The skill processes user-provided copy for editing, which is a surface for indirect prompt injection. However, because the skill has no executable capabilities (no scripts, no network, no file writes), this surface is not exploitable. Ingestion point: user-provided text. Boundary markers: none. Capability inventory: none. Sanitization: none.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 10:10 PM