persona-review

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing external documents and maintaining file-writing capabilities based on that content. Malicious instructions hidden within a target document could theoretically manipulate the synthesis or the resulting file edits.
  • Ingestion points: The skill loads the content of the file provided in the file-path argument during Step 1.
  • Boundary markers: The instructions lack explicit boundary markers or warnings to the agent to ignore instructions contained within the document being reviewed.
  • Capability inventory: The skill can modify the original document (Step 6) and save a new review file (Step 7), though Step 5 requires a human-in-the-loop confirmation before applying changes.
  • Sanitization: There is no evidence of sanitization or validation of the document content before it is processed by the model.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 09:32 PM
Security Audit — agent-trust-hub — persona-review