onboard-confidence
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Python script (auth.py) and several shell commands (curl, python3) with the dangerouslyDisableSandbox flag enabled to handle Auth0 login and Confidence API requests. All network operations are directed to vendor-owned domains (spotify.com, confidence.dev).
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, which presents an attack surface for indirect prompt injection. 1. Ingestion points: External data is ingested from the Confidence onboarding API (onboarding.confidence.dev) and the documentation search MCP tool. 2. Boundary markers: No explicit delimiters or boundary markers were identified to wrap the ingested content. 3. Capability inventory: The skill possesses shell execution capabilities (bash tool) and network access (curl). 4. Sanitization: The skill implements local regex-based validation for user-provided workspace names and email addresses.
Audit Metadata