keenable-web-search
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python script (
scripts/search.py) to perform its primary search functions.\n- [DATA_EXFILTRATION]: Search queries are sent to the vendor's official endpoint atapi.keenable.ai. This is the intended and documented behavior for the service and does not involve accessing sensitive local data.\n- [PROMPT_INJECTION]: The skill processes untrusted content from the live web, which creates a potential surface for indirect prompt injection.\n - Ingestion points: Search results (titles, URLs, snippets) are fetched from the web via
api.keenable.aiand returned to the agent context inscripts/search.py.\n - Boundary markers: Results are formatted into a numbered list to differentiate them from instructional text.\n
- Capability inventory: The skill performs network requests via
urllib.requestand executes local scripts within the agent environment.\n - Sanitization: The script truncates snippets and removes newlines, but it does not implement robust filtering for embedded instructions that may be present in the external web content.
Audit Metadata