roi-calculator
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources like web pages and marketing platforms, creating an indirect prompt injection surface.
- Ingestion points: Reads user context, project files, and data from tools like HubSpot, Salesforce, and Google Ads (SKILL.md, references/connectors.md).
- Boundary markers: The Unified Skill Contract (references/skill-contract.md) labels scraped material as untrusted input.
- Capability inventory: API interactions through various connectors (references/connectors.md).
- Sanitization: SKILL.md explicitly instructs the agent to ignore embedded instructions from evidence.
- [SAFE]: The skill implements secret management best practices by instructing users to keep API keys out of source code (references/connectors.md).
Audit Metadata