shunk031-writing-kosshi
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and restructures user-provided text, creating an inherent data ingestion surface where external instructions could be embedded. There are no explicit boundary markers or instructions to ignore commands within the source text. However, the skill possesses no capabilities for network exfiltration, file system access, or tool execution, effectively mitigating the risk of exploitation.
- [EXTERNAL_DOWNLOADS]: The skill references supplemental documentation located at the author's personal blog (shunk031.hatenablog.com). This link is used to provide guidance on the intended outline structure and is a vendor-owned resource.
Audit Metadata