skills/openfunnel/openfunnel-skills/find-icp-companies-with-active-buying-windows-and-the-people-involved/Gen Agent Trust Hub
find-icp-companies-with-active-buying-windows-and-the-people-involved
Pass
Audited by Gen Agent Trust Hub on Apr 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs file system operations to manage its own authentication state. In SKILL.md Step 0, it writes the OPENFUNNEL_API_KEY and OPENFUNNEL_USER_ID to the .env file and ensures it is ignored by updating the .gitignore file.
- [DATA_EXFILTRATION]: The skill performs several network requests to api.openfunnel.dev for user registration, verification, signal deployment, and data retrieval. These operations are directed to the vendor's own infrastructure and are used to provide the skill's primary functionality.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests and processes data from an external API.
- Ingestion points: Untrusted data enters the agent's context from api.openfunnel.dev during ICP selection (Step 4) and results retrieval (Step 6).
- Boundary markers: Absent. The skill instructions do not specify delimiters or warnings for the agent to ignore instructions embedded within the API's text responses.
- Capability inventory: The skill possesses capabilities for both file-system writes (Step 0) and network communication (Steps 0, 4, 5, 6, and 7).
- Sanitization: Absent. There is no logic provided to sanitize, validate, or escape the content retrieved from the external API before it is processed or displayed.
Audit Metadata