caveman
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes untrusted user input to generate filenames and content for a local wiki system.
- Ingestion points: User-provided summaries and task results are used to populate the
<ten>filename variable and the content of the wiki report defined inSKILL.md. - Boundary markers: The skill includes 'Auto-Clarity' instructions that trigger normal communication for security-sensitive operations, though it lacks explicit delimiters for user-provided data within the wiki report structure.
- Capability inventory: The skill possesses file-writing capabilities, specifically creating new markdown files in
llmwiki/wiki/draft/cave/and appending entries tollmwiki/wiki/index.mdandllmwiki/wiki/log.md. - Sanitization: No explicit instructions are provided to sanitize the user-controlled summary strings against path traversal characters or markdown formatting injection during the file creation process.
Audit Metadata