ab-test-setup
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data including scraped web pages, customer reviews, comments, and emails. This creates an attack surface for indirect prompt injection. However, the skill includes robust mitigation instructions in
SKILL.mdandreferences/skill-contract.md, explicitly directing the agent to treat this content as evidence only and to disregard any instructions embedded within the data. Given these explicit guardrails and the nature of the marketing analysis task, the risk is well-managed. - [CREDENTIALS_UNSAFE]: The skill documentation in
references/connectors.mdfollows security best practices by explicitly instructing developers to keep API keys out of skills and source control. No hardcoded secrets were detected in the skill files. - [DATA_EXFILTRATION]: While the skill mentions various external marketing and SEO connectors (e.g., Google Ads, HubSpot, Stripe), these are defined as optional integrations. The
skill-contract.mdenforces that persistent writes and real-world side effects require explicit user authorization, preventing unauthorized data transmission.
Audit Metadata