map-contacts

Pass

Audited by Gen Agent Trust Hub on Jul 28, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill uses vendor-specific tools (prefixed with swan-) to perform its intended functions. These include swan-update-skill for state management, swan-search-sequences for CRM data, swan-linkedin-social-media-presence for public profile verification, and swan-enrich-contact for data enrichment. All external calls are consistent with the 'swan-gtm' author context.
  • [DATA_EXPOSURE]: The skill accesses CRM, email, and meeting transcript data. However, this access is within the intended scope of a sales mapping tool and relies on the user's previously connected integrations. It does not attempt to exfiltrate this data to unauthorized third parties.
  • [COMMAND_EXECUTION]: The skill instructions involve specific platform tools like FIREFLIES_GET_TRANSCRIPTS, but these are used as data retrieval functions within the agent framework rather than arbitrary shell command execution.
  • [PROMPT_INJECTION]: The instructions contain strict rules and multi-step logic (e.g., 'MUST use the org's defined persona model', 'NEVER enrich every net-new contact'), but these are functional constraints for the AI's behavior rather than attempts to bypass security filters.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 28, 2026, 04:39 PM
Security Audit — agent-trust-hub — map-contacts