handle-linkedin-profile-view-signal
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted data from external sources (LinkedIn profiles via LeadShark webhooks) and interpolates it into downstream systems.
- Ingestion points: Data is received via a webhook payload (documented in
references/webhook-payload-reference-leadshark-new-profile-visit.md), specifically fields likedata.lead.nameanddata.lead.title(LinkedIn headline). - Boundary markers: The skill lacks explicit delimiters or boundary markers when interpolating this external data into CRM notes or Slack messages.
- Capability inventory: The agent has the capability to write to the CRM (Step 6), post to Slack channels (Step 8), and route data to other high-intent or awareness sub-skills (Step 7).
- Sanitization: There is no evidence of sanitization or escaping of the profile content (like headlines) before it is committed to CRM records or chat notifications.
- [EXTERNAL_DOWNLOADS]: The skill integrates with external services including LeadShark (for signal detection) and Apify (as a fallback scraper). These are well-known services for GTM automation and are consistent with the skill's stated purpose.
Audit Metadata