searxng
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill uses standard, well-known Python libraries
httpxandrichfor network operations and CLI formatting. These are specified using standard PEP 723 inline dependency syntax in the main script. - [INDIRECT_PROMPT_INJECTION]: As a web search utility, the skill processes external search results, representing a typical indirect prompt injection surface.
- Ingestion points: External data enters the agent context through search result titles, snippets, and URLs returned by the SearXNG instance in
scripts/searxng.py. - Boundary markers: The skill does not use specific delimiters (like XML tags or 'ignore' markers) around the search results in its output.
- Capability inventory: The skill is capable of network requests to a user-defined URL but does not have file system write access or other high-privilege capabilities.
- Sanitization: Search results are presented to the agent as returned by the engine, with truncation for length but without sanitization of the underlying text.
- [COMMAND_EXECUTION]: The script uses a robust argument parser for handling CLI inputs and does not employ dangerous functions like
eval()oros.system()to process user-supplied search queries.
Audit Metadata