blog-write
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill employs the
Bashtool to runcurlcommands for the purpose of checking the availability of image URLs from external providers. This is a functional requirement for link validation. - [EXTERNAL_DOWNLOADS]: The skill fetches data and media assets from well-known services including Pixabay, Unsplash, and Pexels to populate blog content. These operations are aligned with the skill's primary purpose.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface through its research capabilities.
- Ingestion points: External research statistics and image metadata are ingested from the web using
WebSearchandWebFetchduring Phase 2. - Boundary markers: The instructions do not define specific delimiters to isolate retrieved web content from the agent's internal logic.
- Capability inventory: The skill has access to
Bash,Write, andEdittools, which could be targeted by instructions hidden in web content. - Sanitization: There is no explicit sanitization or validation logic for data retrieved from external sources before it is utilized in the article generation process.
Audit Metadata