handle-demo-request
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from inbound demo requests (form-fills, webhooks), creating a vulnerability surface where malicious instructions could be embedded in the processed data.
- Ingestion points: Lead data, including work email, titles, and company information, is ingested from external demo request forms and webhooks as described in the 'Step 1' and 'Step 2' of SKILL.md.
- Boundary markers: There are no instructions defining boundary markers or explicit 'ignore embedded instructions' warnings to prevent the agent from obeying instructions hidden within lead metadata.
- Capability inventory: The skill has significant capabilities, including creating or updating records in a CRM ({{CRM}}), sending outbound booking or trial emails from {{SALES_SENDER}}, and posting alerts to internal channels ({{ALERT_CHANNEL}}).
- Sanitization: While the skill performs basic gating on email domains, it lacks sanitization or validation for other text fields (e.g., titles or company descriptions) that are used to generate personalized email content or CRM notes, which could lead to prompt injection or content spoofing in outbound communication.
Audit Metadata