skills/gquthier/clawg/native-mcp/Gen Agent Trust Hub

native-mcp

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: The provided file is a documentation resource (SKILL.md) and does not contain any executable scripts or hidden code.
  • [COMMAND_EXECUTION]: The documentation describes how the agent can be configured to execute local subprocesses (such as npx or uvx) to run MCP servers. This is a core feature of the protocol and is controlled by the user through a local configuration file (~/.clawg/config.yaml).
  • [EXTERNAL_DOWNLOADS]: The instructions suggest installing the official mcp library from standard package registries. It also describes connecting to remote HTTP MCP servers, which is part of the intended functionality.
  • [CREDENTIALS_UNSAFE]: The documentation uses illustrative placeholders for API keys (e.g., ghp_..., sk-...). It explicitly details security mechanisms designed to protect user secrets, including filtering the environment variables passed to subprocesses and stripping credentials from error messages shown to the model.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 03:12 PM