emit-pending-review

Pass

Audited by Gen Agent Trust Hub on Jul 16, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides an ingestion surface for untrusted data which is subsequently written to the file system. • Ingestion points: 'Findings' and 'Target' inputs defined in SKILL.md. • Boundary markers: The skill enforces a Markdown schema for the output file but does not provide markers or escaping for the data values themselves. • Capability inventory: File writing to the '.pending-reviews/' directory. • Sanitization: No content sanitization or instructions to ignore embedded instructions are present.
  • [SAFE]: The skill implements strict side-effect limitations and input validation. It explicitly forbids writing to sensitive directories (e.g., '.pending-decisions/') and includes clear ownership boundaries that prevent the skill from executing or routing the findings it records.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 16, 2026, 08:33 PM
Security Audit — agent-trust-hub — emit-pending-review