io.github.CSOAI-ORG/agent-delegation-mcp

The materials explicitly state that no keys or environment variables are required. There is no indication that the tool asks for API keys, OAuth tokens, or local sensitive credentials, so the credential exposure surface appears limited based on known facts.

System checks confirm outbound network access to the remote endpoint api.meok.ai. For an MCP tool, connecting to its declared service endpoint is a normal capability, but the available materials do not specify what user inputs, context, or results may be transmitted, so data egress scope should be reviewed.

System checks indicate the tool has code-execution capability, meaning it can start a local service or run code on the host. This is a common operational pattern for MCP tools; there is no concrete sign here of excessive system-level privilege requests, but its actual runtime behavior should still be observed in an isolated environment.

The available materials provide no README or permission details, so it is unclear what local files, session data, or other resources it can read or write. There is no explicit over-permissioning red flag, but given that it executes code as an MCP tool, it should be deployed with minimal data access by default.

Positive signals include an official registry source, open-source code on GitHub, and updates within the past year; these materially reduce supply-chain risk. Caution remains because the README is absent, the license is undeclared, and community adoption is very low (0 stars), so while the code is auditable, external validation is limited.