Safe AI Agent MCP
A free, read-only MCP server providing three diagnostic tools for AI agents in production: auditing API-key privileges to flag over-privileged keys, checking code snippets for idempotency patterns to prevent duplicate execution, and explaining production guard patterns like idempotency keys and circuit breakers. It makes no network calls and refuses inputs resembling real secrets.
Package and access
Local package with source-attributed installation metadata. Koskamo does not execute packages during review.
What you can do with it
- Audit API-key privileges to identify over-privileged keys
- Check code snippets for idempotency patterns
- Explain production AI-agent guard patterns
Source and maintenance
- Maintenance checked
Packages and installation
Source-provided package data. Koskamo has not executed or live-probed this package.
safe-agent-mcp 0.1.0
{
"version": "0.1.0",
"transport": "stdio",
"identifier": "safe-agent-mcp",
"registryType": "pypi"
}Tools
Reported by the attributed source; not independently observed by Koskamo.
Connect from your client
Each snippet configures the same source-provided package; replace placeholders before use.
Cursor
{
"mcpServers": {
"safe-ai-agent-mcp": {
"command": "uvx",
"args": [
"safe-agent-mcp==0.1.0"
]
}
}
}Cline
{
"mcpServers": {
"safe-ai-agent-mcp": {
"command": "uvx",
"args": [
"safe-agent-mcp==0.1.0"
]
}
}
}VS Code
{
"servers": {
"safe-ai-agent-mcp": {
"type": "stdio",
"command": "uvx",
"args": [
"safe-agent-mcp==0.1.0"
]
}
}
}Claude Desktop
{
"mcpServers": {
"safe-ai-agent-mcp": {
"command": "uvx",
"args": [
"safe-agent-mcp==0.1.0"
]
}
}
}Let your agent find its tools
Find servers like Safe AI Agent MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.