growth-diagnosis

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a robust 'Completion Gate' and 'Gate The Intake' process, ensuring the agent does not fabricate data or diagnostics when information is missing, defaulting instead to a 'triage' mode.
  • [SAFE]: The 'Evidence And Routing' reference (references/evidence-and-routing.md) and 'Funnel Contract' (references/modules/funnel-analysis/references/funnel-contract.md) establish strict data privacy rules, explicitly instructing the agent to minimize or redact credentials, tokens, cookies, personal data, and sensitive URLs.
  • [SAFE]: The skill defines a clear 'Authorization Boundary' across all modules, stating that diagnosis is read-only by default and requires separate, explicit authorization for any external system changes, campaign publications, or spend adjustments.
  • [SAFE]: The capability assessment instructions (references/modules/growth-maturity-assessment/SKILL.md) include specific prohibitions against identifying, ranking, or surveilling individuals, focusing the analysis on work systems and aggregate data to protect employee privacy.
  • [SAFE]: The skill uses a structured JSON schema (assets/protocol.schema.json) for its final output, which restricts the agent's ability to inject unauthorized data fields or formats through the 'additionalProperties: false' constraint.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 01:11 AM