film-first-ad

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: An indirect prompt injection surface was identified in the skill's data processing logic.
  • Ingestion points: The skill takes production details and file paths through user-influenced input which are then passed to a local script via command-line arguments.
  • Boundary markers: The skill lacks markers to distinguish between system instructions and user-provided data, potentially allowing embedded instructions to be executed by the agent.
  • Capability inventory: The Python script scripts/generate_call_sheet.py contains file-writing functionality.
  • Sanitization: No validation or sanitization is performed on the output file path or the content being written, creating a risk of arbitrary file overwrite.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 08:06 AM
Security Audit — agent-trust-hub — film-first-ad