humanizer-cn
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, potentially untrusted text to identify AI patterns. Because the agent is granted file-system capabilities (Read, Write, Edit), it is theoretically vulnerable to indirect prompt injection where instructions hidden in the text could influence the agent's actions.\n
- Ingestion points: Untrusted Chinese text provided for 'humanization' analysis (SKILL.md).\n
- Boundary markers: The skill lacks explicit delimiters or instructions to ignore commands embedded within the text to be processed.\n
- Capability inventory: The skill uses tools for reading and writing to the file system (SKILL.md frontmatter).\n
- Sanitization: No validation or sanitization of the input text is specified to prevent command execution.
Audit Metadata