Hunch runs on your own machine — the client starts it, so there is no endpoint to ping. 410 installs a week from npm. Last commit 11 Oct 2026.
Shared records for AI agents: engineering decisions, bug history, rules, and work state through MCP.
We read the source, 4 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.
const run = spawnSync(process.execPath, [
vm.runInNewContext(readFileSync(postsFile, "utf8"), sandbox);
const targetDir = path.join(repoRoot, "site", locale);
… page keys PR- and merge-influenceable), yet every key drives file write/delete operations. Traversing keys like '../../id_rsa' enabled arbitrary file deletion without per-site validation. Touches invariant(s): con_e04226bd05 (The MCP server must stay client-agnostic: all assista…
document.getElementById("grid").innerHTML = rest.map((p) => `
for (const [k, v] of Object.entries(process.env)) if (!k.startsWith("GIT_")) env[k] = v;
Is this your server and something here is wrong? Tell us — corrections are free and do not require a plan.
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.
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 hunch -- npx -y @davesheffer/hunch
{
"mcpServers": {
"hunch": {
"args": [
"-y",
"@davesheffer/hunch"
],
"command": "npx"
}
}
}
[mcp_servers.hunch]
command = "npx"
args = ["-y", "@davesheffer/hunch"]
{
"mcpServers": {
"hunch": {
"args": [
"-y",
"@davesheffer/hunch"
],
"command": "npx"
}
}
}
{
"mcpServers": {
"hunch": {
"args": [
"-y",
"@davesheffer/hunch"
],
"command": "npx"
}
}
}
Cross-harness communication and shared work-state for AI coding agents.
Session memory for AI coding agents: records decisions and dead ends, flags what went stale.
Shared memory for coding agents and their teams: docs, epics, tasks and decisions over MCP.
Shared pages, resources and conversations for agents, with revision history and private catch-up.
Shared project memory for Claude Code, Codex and Cursor: decisions and corrections through MCP.
Memory for coding agents: the decisions, the dead ends, and where the last session stopped.
Shared project memory for AI coding agents: decisions, lessons, risks and tasks in one graph.
Memory, docs, and repo context layer for engineering agents.
Answers built from our own checks of this server.