inkos-long-story-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external story materials (files, PDFs, URLs) via the
ingest_materialfunction. This creates a surface for indirect prompt injection where a processed novel could contain hidden instructions intended to influence the agent. However, this is a standard risk for analysis skills and is mitigated by the skill's specific focus on semantic craft extraction rather than execution of content. The severity is low because the skill lacks capabilities for network exfiltration or direct file system modifications that could be exploited by such an injection.
Audit Metadata