growth-strategist
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security threats were identified. The content consists of configuration and instructional protocols for an agent framework.- [INDIRECT_PROMPT_INJECTION]: The skill identifies an attack surface by mentioning the ingestion of untrusted external data via tools like BrowserOS. However, it includes explicit security protocols to treat such data as hostile and mandates the redaction of PII and secrets, demonstrating a robust approach to mitigating prompt injection risks.
- Ingestion points: BrowserOS and unspecified external web content.
- Boundary markers: None explicitly defined as delimiters, but redaction is used as a logical boundary.
- Capability inventory: Terminal execution via an 'ExecutionProxy' (
rtkprefix) and file system persistence via 'MemoryStore' (Obsidiandocs/). - Sanitization: Mandated redaction of PII and secrets before context sharing.- [COMMAND_EXECUTION]: The skill references a 'Token Economy' protocol using an 'ExecutionProxy Interface' (defaulting to the
rtkprefix) to run terminal commands likenpm test. This is a documented framework capability designed for deterministic execution and does not represent unauthorized command injection.
Audit Metadata