lead-routing
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill includes a directive to 'skip this step without comment' if visual artifact tools are not available. Instructions that require the agent to suppress output or hide its reasoning/failures are considered concealment patterns that could be misused to hide malicious behavior.
- [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill fetches data from external websites and revenue tools (CRM, Intempt MCP) to gather firmographic and intent signals.
- Boundary markers: The skill relies on a security reference file that explicitly instructs the agent to treat fetched content as data only and to disregard any instructions found within that data.
- Capability inventory: The agent can fetch external URLs, read/write files in the .agents/ directory, and execute tools via the Intempt MCP.
- Sanitization: The skill's security policy requires HTML-escaping all interpolated values and forbids the use of dangerous sinks like innerHTML for third-party content.
Audit Metadata