polpo-integrations
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious behavior detected. The provided files consist of architectural documentation and configuration schemas that guide developers in creating secure agent integrations.
- [SAFE]: Secure Credential Management. The skill promotes a 'security by design' approach by instructing that connection IDs and credentials must not be controlled by model arguments. Instead, it utilizes encrypted 'Connections' and host-owned 'ToolInvocationContext' to resolve identity and authorization at runtime.
- [SAFE]: Execution Sandboxing. The documentation references a platform capability to mark tools with
requiresSandbox, which applies restrictive filesystem and shell policies during tool execution. - [SAFE]: Input Validation and Bindings. The skill uses hidden bindings and
bindingsSchema(TypeBox) to ensure that sensitive parameters are mapped directly from trusted host metadata rather than being exposed to or manipulated by the LLM.
Audit Metadata