finding-real-estate-professionals-on-twitter
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external Twitter profiles (bios, usernames). As with any skill processing untrusted external content, there is an inherent surface for indirect prompt injection if the agent interprets text from these profiles as instructions.
- Ingestion points: Twitter user data (bio, username, location, website) retrieved via the Apify actor.
- Boundary markers: Absent in the description of how the agent should handle the data.
- Capability inventory: Shell execution (node, curl) and Apify MCP tool access.
- Sanitization: Not explicitly defined in the instructions.
- [DYNAMIC_EXECUTION]: The
customMapFunctioninput parameter allows for providing a JavaScript string to transform output data. This is a standard feature of the underlying Apify actor used for data processing, though it represents a surface for dynamic code execution if untrusted code is supplied to this parameter. - [COMMAND_EXECUTION]: The documentation includes shell commands for executing local scripts (
node scripts/run_actor.js) and making API requests (curl). These are standard operational instructions for interacting with the service.
Audit Metadata