kelly-crm
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from emails and meeting notes to draft messages. This is mitigated by a mandatory human review workflow where all items start in a 'needs_review' state. Evidence: Ingestion points include Busabase records (busabase-provider.js); Boundary markers include the review workflow (SKILL.md); Capability inventory includes Busabase writes and handoff metadata (execute_decisions.mjs); Sanitization is handled by HTML escaping in the UI (crm-views.js).
- [DATA_EXPOSURE_AND_EXFILTRATION]: Credential management follows best practices by using environment variables (BUSABASE_API_KEY). The skill maintains a strict boundary against direct external communication, delegating side effects to other skills only after user approval.
- [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill uses established Node.js dependencies and provides administrative scripts for setup that interact exclusively with the Busabase API using trusted credentials.
Audit Metadata