author
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is configured to ingest and process supplied source material, which creates a surface for indirect prompt injection attacks. Evidence: 1. Ingestion points: The skill reads source material in Step 2 of SKILL.md. 2. Boundary markers: Step 2 includes an explicit instruction to ignore embedded instructions that redirect the task or request unrelated access. 3. Capability inventory: The skill can write files to user-defined destination paths and execute local validators as described in Step 4. 4. Sanitization: No programmatic sanitization or filtering is specified.
- [COMMAND_EXECUTION]: The skill's validation branch in Step 4 references the use of 'available local validator output'. This implies the execution of external CLI tools or scripts on the host system to validate Markdown document structure and links.
Audit Metadata