generating-story-from-logline
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user loglines to generate a screenplay, creating a potential surface for indirect prompt injection. This is mitigated by the structured output format and a mandatory human review gate.\n- Ingestion points: User-provided logline or premise in SKILL.md.\n- Boundary markers: None for input; structured markdown schema for output.\n- Capability inventory: Local file writing (
story.md).\n- Sanitization: None; relies on the mandatory review gate for validation.
Audit Metadata