Signoff CLI runs on your own machine — the client starts it, so there is no endpoint to ping. 2 649 installs a week from npm. Last commit 8 Nov 2026.
Five gates that make an agent's 'done' provable: run them, read the verdict, audit the record.
What the code does
We read the source, 7 h ago · rules 995398eba69f
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.
postinstall: node bin/signoff.js ensure --from-install --quiet
const r = spawnSync(python(), [file, ...args], { stdio: "inherit" });
proc = subprocess.run(cmd, shell=True, capture_output=True, text=True, errors="replace")
proc = subprocess.run(cmd, shell=True, capture_output=True, text=True, errors="replace")
return path.isAbsolute(raw) ? raw : path.join(root, raw);
for (const n of ["id_ed25519", "id_rsa", "id_ecdsa"]) out.push(path.join(home, ".ssh", n));
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 signoff-cli -- npx -y signoff-cli{
"mcpServers": {
"signoff-cli": {
"command": "npx",
"args": [
"-y",
"signoff-cli"
]
}
}
}[mcp_servers.signoff-cli]
command = "npx"
args = ["-y", "signoff-cli"]{
"mcpServers": {
"signoff-cli": {
"command": "npx",
"args": [
"-y",
"signoff-cli"
]
}
}
}{
"servers": {
"signoff-cli": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"signoff-cli"
]
}
}
}Alternatives to Signoff CLI
same job, measured the same wayFive gates that make an agent's 'done' provable: run them, read the verdict, audit the record.
Make a coding agent prove it: run the repo's real tests, typecheck and lint before it says done.
Fork-replay one recorded span from a failed agent run and get an attribution verdict.
Append-only, signed, on-chain-verifiable agent action log — record and audit what agents did.
Hash-chained audit ledger for AI agent work — record, verify and replay what the agent did.
Read tables, figures and dates out of a PDF, Word file, CSV or paste, with a verdict on the reading.
Record x402 agent purchases as reconciled, auditable receipts; verify any receipt on the ledger.
Record an agent's decisions with reasons anyone can re-execute offline — verify recomputes them.
Signoff CLI — questions
Answers built from our own checks of this server.