entrepreneurial-demand-generation
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external business data which could theoretically contain malicious instructions designed to influence the agent's behavior.
- Ingestion points: The skill is designed to ingest data from client intake, primary research, CRM records, analytics, sales records, and customer interviews (defined in
SKILL.mdunder the Inputs tables). - Boundary markers: The skill contains a 'Capability Contract' section that explicitly restricts the agent from launching campaigns, contacting customers, spending budget, or altering CRM data without explicit authority, providing a policy-level boundary.
- Capability inventory: The skill itself does not contain scripts or commands that execute shell operations, network exfiltration, or privileged file system writes.
- Sanitization: No explicit technical sanitization or validation of the input data is described, relying instead on the established decision rules and capability contract.
Audit Metadata