Serve JevGate to coding agents over MCP

.github/homebrew-formula.sh

@@ -48,7 +48,7 @@ class Jevgate < Formula
     generate_completions_from_executable(bin/"jevgate", "completions")
     man1.mkpath
     (man1/"jevgate.1").write Utils.safe_popen_read(bin/"jevgate", "man")
-    %w[auth check baseline rules init serve completions man].each do |command|
+    %w[auth check baseline rules init serve mcp completions man].each do |command|
       (man1/"jevgate-#{command}.1").write Utils.safe_popen_read(bin/"jevgate", "man", command)
     end
   end

CHANGELOG.md

@@ -4,6 +4,8 @@ Notable changes to JevGate. Versions follow [Semantic Versioning](https://semver
 
 ## [Unreleased]
 
+- `jevgate mcp` runs a Model Context Protocol server on stdin and stdout, so coding agents can call JevGate as a tool: `jevgate_check` runs a check and returns its findings (an incomplete run is a tool error), `jevgate_findings` reads the last report, and `jevgate_rules` lists the rules.
+
 ## [0.18.0] - 2026-09-25
 
 - Documentation site: https://tech-byte-frontier.github.io/jevgate/, with guides, troubleshooting, a page on coding agents, and rules, configuration and command-line references generated from the binary. It is published with each release.

site/generate.py

@@ -13,7 +13,7 @@
 import sys
 from pathlib import Path
 
-COMMANDS = ["auth", "check", "baseline", "rules", "init", "completions", "man", "serve"]
+COMMANDS = ["auth", "check", "baseline", "rules", "init", "completions", "man", "serve", "mcp"]
 GROUPS = {
     "maintainability": "On by default.",
     "tests": "On by default; judged with `--include-tests` or `include_tests = true`.",

site/src/coding-agents.md

@@ -19,6 +19,28 @@ for a `consider`, fix it or say why the code should stay as it is.
 | 1 | The gate failed: act on the findings listed |
 | 2 | The run could not finish (no key, provider rejection, request budget); report it, don't treat it as a pass |
 
+## As an MCP server
+
+`jevgate mcp` is a [Model Context Protocol](https://modelcontextprotocol.io) server on stdin and stdout, so an agent can call JevGate as a tool instead of running a shell command. Register it, started in the repository:
+
+```sh
+claude mcp add jevgate -- jevgate mcp          # Claude Code
+```
+
+```json
+{"mcpServers": {"jevgate": {"command": "jevgate", "args": ["mcp"]}}}
+```
+
+The second form is for clients configured with JSON, such as Cursor. The server offers three tools:
+
+| Tool | What it does |
+|---|---|
+| `jevgate_check` | Runs `jevgate check` in the repository with `base`, `paths`, `rules`, `include_tests`, `dry_run` or `verbose`, and returns the ranked findings. An incomplete run (exit 2) is a tool error, never a pass |
+| `jevgate_findings` | Reads the last report's findings, optionally under one path, without running anything |
+| `jevgate_rules` | Lists every rule with the question it asks |
+
+A check runs as a child process with the repository's `jevgate.toml` and key, so the tool reviews exactly what the command line would.
+
 ## Structured output
 
 `--format json` prints the full report: every file, finding, raw answer and probability, and the gate. The same report is always written to `.jevgate/latest.json`, whatever the output format, so an agent can run the check once and read the details after. `jevgate check --help` explains its fields.

src/command.rs

@@ -3,7 +3,7 @@
 use crate::{
     auth, baseline, cancellation, catalog, config,
     config::ConfigContext,
-    init, manual,
+    init, manual, mcp,
     options::{self, JevCommand},
     output, revision, server,
 };
@@ -15,6 +15,7 @@ pub fn run(command: JevCommand) -> Result<u8> {
         JevCommand::Completions { shell } => manual::completions(shell).map(|()| 0),
         JevCommand::Man { command } => manual::man(command.as_deref()).map(|()| 0),
         JevCommand::Init { force } => init(force),
+        JevCommand::Mcp => mcp::run().map(|()| 0),
         command => configured(command),
     }
 }
@@ -44,7 +45,8 @@ fn configured(command: JevCommand) -> Result<u8> {
         JevCommand::Auth { .. }
         | JevCommand::Init { .. }
         | JevCommand::Completions { .. }
-        | JevCommand::Man { .. } => {
+        | JevCommand::Man { .. }
+        | JevCommand::Mcp => {
             unreachable!("handled before repository configuration")
         }
         JevCommand::Check(mut args) => {

src/main.rs

@@ -43,6 +43,7 @@ mod inventory;
 mod line_ranges;
 mod locations;
 mod manual;
+mod mcp;
 mod options;
 mod output;
 mod packages;