scout
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its processing of untrusted web data.\n
- Ingestion points: External content is downloaded from user-supplied links or search engines via the
web-fetchandweb-searchintegrations as described inSKILL.mdunder Step 3 (Content Fetch & Analysis).\n - Boundary markers: Absent. There are no markdown or XML delimiters specified to isolate the retrieved web text from the agent instructions.\n
- Capability inventory: The skill reads profile configuration files (
MY-PROFILE.mdandMY-INTERESTS.md) and passes verified categories to the/url-dumpcapability. It does not use arbitrary shell commands or file write operations.\n - Sanitization: Absent. No input validation, escaping, or heuristic parsing is performed on the text downloaded from the web.
Audit Metadata