linkedin-job-post-to-buyer-pain-map
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted job descriptions from external sources and incorporates them into LLM prompts to perform scoring and pain analysis.
- Ingestion points: Job description text collected in Step 2c of
SKILL.md. - Boundary markers: The instructions lack clear delimiters (e.g., XML tags or triple quotes) when passing user-provided job text to the Gemini API in Steps 4 and 5.
- Capability inventory: The skill has the ability to write to the local filesystem (
mkdir,catin Step 8) and make network requests viacurlto the Gemini API. - Sanitization: No validation or sanitization is performed on the job post text before it is sent to the LLM.
Audit Metadata