directing-protagonist-endangered
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues were detected. The skill consists of Markdown-based guidelines and a JSON file for testing triggering conditions, neither of which contain executable code or dangerous commands.
- [PROMPT_INJECTION]: The instructions are strictly focused on narrative theory (Mackendrick's film directing principles) and do not contain any patterns intended to bypass safety filters or override system instructions.
- [DATA_EXFILTRATION]: There are no commands or functions that access sensitive files or perform network requests. The skill operates entirely within the textual context provided to the agent.
- [REMOTE_CODE_EXECUTION]: The skill does not include any scripts, package dependencies, or remote download instructions.
- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests user-provided story descriptions (Ingestion points: SKILL.md Execution steps). No boundary markers are present. Capability inventory shows no subprocess calls, network ops, or file writes. No sanitization is performed, but no dangerous tools are exposed.
Audit Metadata