chief-customer-officer-advisor
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's workflows ingest untrusted data from user-provided JSON files which could potentially influence agent behavior if they contain malicious instructions. * Ingestion points: User-provided files cx_state.json, at_risk_accounts.json, and voc_state.json are processed by the included Python scripts. * Boundary markers: Absent; the agent is instructed to process script outputs directly into strategic documents without explicit warnings to ignore embedded commands. * Capability inventory: The skill utilizes local Python scripts (cx_maturity_scorer.py, churn_intervention_planner.py, voc_program_designer.py) to transform data. * Sanitization: The scripts do not perform validation or sanitization on input text before including it in the generated reports.
Audit Metadata