golden-age-book-writing
Pass
Audited by Gen Agent Trust Hub on Apr 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill focuses on providing instructions for prose generation and iterative refinement based on a defined 'author council' methodology.
- [DATA_EXFILTRATION]: The skill incorporates a research protocol that utilizes web searches to gather domain-specific information and supporting evidence for writing. This is a standard functional component for research-driven content creation and does not target sensitive user data or involve unauthorized network transmissions.
- [PROMPT_INJECTION]: The instructions establish stylistic guidelines and quality rubrics. There are no attempts to bypass safety filters, override system instructions, or extract internal configuration data. The skill includes a research protocol (Phase 1 Domain Research in SKILL.md) that ingests external data from web searches; while this creates an indirect prompt injection surface, the absence of high-privilege capabilities such as arbitrary command execution or sensitive file access minimizes the risk to a safe level.
Audit Metadata