novel-strict-review
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it is designed to ingest and analyze external, untrusted content (novel chapters).
- Ingestion points: The skill reads external text from paths such as
工作区/第X章VX/正文.mdas defined inSKILL.md. - Boundary markers: The instructions do not define explicit boundary markers or provide instructions to the AI to ignore commands that might be embedded within the novel text.
- Capability inventory: The skill has the capability to read local files (chapters) and write new files (review reports) within the designated workspace directory.
- Sanitization: There is no evidence of sanitization, filtering, or escaping of the ingested text before it is processed by the agent.
Audit Metadata