event-prospecting-pipeline
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow that ingests untrusted data from external sources, creating a potential surface for indirect prompt injection.
- Ingestion points: The skill fetches person names, bios, and social media profiles from external sources via
luma-event-attendeesandconference-speaker-scraperas described inSKILL.mdStep 1. - Boundary markers: There are no explicit instructions or delimiters defined in the playbook to separate untrusted external bios from agent instructions.
- Capability inventory: The skill has the capability to send personalized emails via the
cold-email-outreachskill and the AgentMail API (agentmail.dev) as described inSKILL.mdStep 7. - Sanitization: The skill does not specify any sanitization or filtering of the external profile data before it is used to qualify leads or generate outreach content.
- [EXTERNAL_DOWNLOADS]: The skill mentions using the AgentMail API (
agentmail.dev) for launching outreach. This external service is consistent with the skill's stated purpose of providing an end-to-end lead generation and outreach pipeline.
Audit Metadata