blog-cannibalization
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from blog files (.md, .mdx, .html).
- Ingestion points: Local content files scanned in the target directory as specified in SKILL.md.
- Boundary markers: The skill contains a clear directive for the agent to ignore any instructions embedded in the blog content.
- Capability inventory: Limited to file reading, semantic analysis, and API reporting; it does not perform file writes or modifications.
- Sanitization: Implements deterministic normalization and tokenization of extracted text.
- [COMMAND_EXECUTION]: Employs standard file system tools (Glob, Grep) and a dedicated local CLI wrapper to interface with the DataForSEO API.
- [DATA_EXFILTRATION]: Communicates with the DataForSEO API (api.dataforseo.com), a well-known service. The skill follows best practices by using environment variables for credentials and obtaining user approval for API request volume and cost before proceeding.
Audit Metadata