gutsy runs on your own machine — the client starts it, so there is no endpoint to ping. 1 020 installs a week from pypi. Last commit 7 Oct 2026.
Local gut check before risky agent actions: calibrated safe/intended probabilities, no API calls.
What the code does
We read the source, 3 d ago · rules ccca72095570
What this server is able to do. For an MCP server this is often the job itself — a terminal server runs commands because that is what it is for. Listed so you know what you are plugging in, not as an accusation.
_server_proc = subprocess.Popen(
Is this your server and something here is wrong? Tell us — corrections are free and do not require a plan.
This code can reach further than it looks
We found places where it runs commands, builds paths or queries from values it is given. None of that is a flaw by itself — it becomes one when the code changes, and code changes quietly between releases. We re-read it on every one.
Connect this server
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add gutsy -- uvx gutsy-mcp{
"mcpServers": {
"gutsy": {
"args": [
"gutsy-mcp"
],
"command": "uvx"
}
}
}[mcp_servers.gutsy]
command = "uvx"
args = ["gutsy-mcp"]{
"mcpServers": {
"gutsy": {
"args": [
"gutsy-mcp"
],
"command": "uvx"
}
}
}{
"mcpServers": {
"gutsy": {
"args": [
"gutsy-mcp"
],
"command": "uvx"
}
}
}This one needs environment variables set before it will start:
GUTSY_URL (URL of the local gutsy-inference server), GUTSY_INFERENCE_DIR (Directory with models.json; if set, gutsy-mcp starts gutsy-inference itself), GUTSY_PROCEED_MIN (Minimum P(proceed) and P(intended) to recommend proceeding), GUTSY_BLOCK_MIN (Minimum P(block) to recommend blocking).
The author declared them in the registry entry; get the values from the project itself.
Alternatives to gutsy
same job, measured the same wayCalibrated decisions for agents: choice, yes/no and score questions answered with probabilities.
Calibrated advice for coding agents: typed questions in, probabilities out. Apple Silicon.
Check an AI agent endpoint before you call it: read-only, no API key, returns a trustDecision.
Local fit-check and safe integration guidance for non-custodial Base USDC agent transactions.
MCP server over TypeSafe Jev: typed ask, rank and check judgments return calibrated probabilities.
Calibrated judgments for text: yes/no probabilities, picks from your options, or scores.
logi Agent Approval Gate — human approval on the phone before high-risk agent actions.
MCP server for Jev (TypeSafe System One): typed decisions with calibrated probabilities.
gutsy — questions
Answers built from our own checks of this server.