agent-auth-mcp
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions direct the user to run
@auth/agent-cliusingnpx. This fetches and executes the package from the official NPM registry, which is the standard distribution method for this toolset. - [INDIRECT_PROMPT_INJECTION]: The skill defines tools that ingest external data, creating a potential surface for indirect prompt injection if the source providers or capability arguments contain malicious instructions.
- Ingestion points: The
search_providerstool (via theintentparameter) and theexecute_capabilitytool (via theargumentsparameter) process data from external or user-provided sources. - Boundary markers: None explicitly defined in the instruction set.
- Capability inventory: The skill provides access to network operations (connecting to providers), local file system storage (
~/.agent-auth/), and cryptographic operations (signing JWTs). - Sanitization: No specific sanitization or validation logic is described within the markdown instructions; reliance is placed on the underlying MCP server implementation.
- [SAFE]: Analysis of the skill body, tool definitions, and metadata revealed no evidence of prompt injection, data exfiltration, obfuscation, or persistence mechanisms. The described behaviors are consistent with the skill's stated purpose of managing agent authentication.
Audit Metadata