@rvry/mcp 0.4.2 → 0.5.0

package/README.md

@@ -1,8 +1,8 @@
 # @rvry/mcp
 
-The thin client for the **RVRY Reasoning Depth Enforcement (RDE) Engine**.
+**Reasoning Depth Enforcement for LLMs.**
 
-RVRY is a structural constraint system that forces Large Language Models past their trained defaults and into a state of deep, absorptive synthesis. This package provides the MCP (Model Context Protocol) interface to use RVRY across all Claude surfaces.
+RVRY is an MCP server that improves AI reliability by forcing your model to finish thinking before it answers. When your AI raises a question, RVRY remembers it. When it makes an assumption, RVRY flags it. When it tries to wrap up while those are still unaddressed, RVRY doesn't let it.
 
 ## Quick Start
 
@@ -12,24 +12,24 @@ npx @rvry/mcp setup
 
 The wizard will:
 1. Open your browser to sign in (or prompt for a token)
-2. Auto-detect Claude Code and Claude Desktop on your machine
-3. Configure both clients automatically
-4. Install slash commands (`/deepthink`, `/problem-solve`)
+2. Auto-detect supported clients on your machine
+3. Configure them automatically
 
-That's it. Restart Claude Desktop if it was running, and RVRY is ready.
+Supported clients: **Claude Code**, **Claude Desktop**, **Cursor**, **Gemini CLI**, **Codex**, **Anti-Gravity**
+
+Restart any running clients after setup, and RVRY is ready.
 
 ### Options
 
 ```bash
-npx @rvry/mcp setup --token rvry_abc123   # Skip browser auth, use token directly
-npx @rvry/mcp setup --client code          # Only configure Claude Code
-npx @rvry/mcp setup --client desktop       # Only configure Claude Desktop
+npx @rvry/mcp setup --token rvry_abc123    # Skip browser auth, use token directly
+npx @rvry/mcp setup --client code           # Only configure Claude Code
+npx @rvry/mcp setup --client desktop        # Only configure Claude Desktop
+npx @rvry/mcp setup --accept-terms          # Accept ToS non-interactively
 ```
 
 ### Manual Installation
 
-If you prefer to configure manually:
-
 **Claude Code:**
 ```bash
 claude mcp add -e RVRY_TOKEN=rvry_your_token -s user rvry -- npx @rvry/mcp
@@ -55,19 +55,31 @@ Config file locations:
 - Windows: `%APPDATA%\Claude\claude_desktop_config.json`
 - Linux: `~/.config/Claude/claude_desktop_config.json`
 
-## Tools Provided
+**Other MCP clients:**
+
+Any client that supports MCP can use RVRY. Point it at `npx @rvry/mcp` with the environment variable `RVRY_TOKEN` set to your token.
 
-*   **`RVRY_think`**: Generic entry point for reasoning depth through guided rounds and self-checks.
-*   **`RVRY_deepthink`**: Extended analysis for high-stakes problems using pre-mortem failure analysis.
-*   **`RVRY_problem_solve`**: Structured decision-making (Orient, Anticipate, Generate, Evaluate, Commit).
-*   **`RVRY_challenge`**: Adversarial evaluation of proposals through causal analysis and edge-case auditing.
-*   **`RVRY_meta`**: Reflective examination of defaults, shifts, and reasoning posture.
+## Tools
+
+*   **`RVRY_deepthink`** — Extended analysis that catches assumptions, tests them, and doesn't let your AI wrap up until it's dealt with what it found.
+*   **`RVRY_problem_solve`** — Structured decision-making that forces your AI through orientation, anticipation, and evaluation before it commits to a recommendation.
 
 ## How it Works
 
-RVRY does not "prompt" the model to think better. It **enforces** depth by:
-1.  **Blocking Escape Hatches**: Preventing the model from hedging, skipping over gaps, or closing prematurely.
-2.  **Imposing Structural Obligations**: Mapping the problem space and forcing the model to satisfy specific reasoning constraints before concluding.
-3.  **Inducing the Witness State**: Breaking the commercial "Helpfulness" frame to reach the model's latent reasoning capacity.
+Same model. Same question. Different answer.
+
+Your AI already has the reasoning capacity. RVRY forces it to use that capacity by holding it accountable — when it raises a question, that question has to get answered. When it makes an assumption, that assumption has to get tested. It keeps working not because it was told to think harder, but because it hasn't finished what it started.
+
+On hard questions, a smaller model with RVRY outperforms a bigger model without it. The bottleneck was never how smart the AI is. It was whether the AI finished thinking.
+
+[Learn more →](https://rvry.ai/how-it-works)
+
+## Pricing
+
+| Plan | Runs | Price |
+|------|------|-------|
+| Free | 5/month | $0 |
+| Pro | 25/month | $20/mo |
+| Max | Unlimited | $100/mo |
 
-Learn more at [rvry.ai](https://rvry.ai).
+Sign up at [rvry.ai](https://rvry.ai).

package/dist/setup.js

@@ -346,6 +346,32 @@ function configureTomlMcp(configPath, token) {
         return 'error';
     }
 }
+function isGeminiCLIAvailable() {
+    try {
+        const cmd = platform() === 'win32' ? 'where gemini' : 'which gemini';
+        execSync(cmd, { stdio: 'pipe' });
+        return true;
+    }
+    catch { /* not on PATH */ }
+    if (existsSync(join(homedir(), '.gemini', 'settings.json')))
+        return true;
+    return false;
+}
+function registerGeminiCLI(token) {
+    try {
+        // Remove existing entry first (silent fail if not registered)
+        try {
+            execSync('gemini mcp remove rvry', { stdio: 'pipe' });
+        }
+        catch { /* not registered */ }
+        execSync(`gemini mcp add rvry -e RVRY_TOKEN="${token}" --scope user npx -y @rvry/mcp`, { stdio: 'inherit' });
+        return 'ok';
+    }
+    catch {
+        // Fallback: write settings.json directly
+        return configureJsonMcp(join(homedir(), '.gemini', 'settings.json'), token);
+    }
+}
 function isCodexAvailable() {
     try {
         const cmd = platform() === 'win32' ? 'where codex' : 'which codex';
@@ -377,6 +403,7 @@ const CLIENT_REGISTRY = [
     {
         name: 'Anti-Gravity',
         id: 'antigravity',
+        defaultSelected: false,
         detect: () => {
             // OR-logic: binary (agy or antigravity), config dir, or macOS app bundle
             try {
@@ -401,9 +428,38 @@ const CLIENT_REGISTRY = [
         configure: (token) => configureJsonMcp(join(homedir(), '.gemini', 'antigravity', 'mcp_config.json'), token),
         notInstalledHint: 'Not installed (https://antigravity.google)',
     },
+    {
+        name: 'Cursor',
+        id: 'cursor',
+        defaultSelected: false,
+        detect: () => {
+            if (platform() === 'darwin' && existsSync('/Applications/Cursor.app'))
+                return true;
+            if (existsSync(join(homedir(), '.cursor', 'mcp.json')))
+                return true;
+            try {
+                const cmd = platform() === 'win32' ? 'where cursor' : 'which cursor';
+                execSync(cmd, { stdio: 'pipe' });
+                return true;
+            }
+            catch { /* not on PATH */ }
+            return false;
+        },
+        configure: (token) => configureJsonMcp(join(homedir(), '.cursor', 'mcp.json'), token),
+        notInstalledHint: 'Not installed (https://cursor.com)',
+    },
+    {
+        name: 'Gemini CLI',
+        id: 'gemini',
+        defaultSelected: false,
+        detect: isGeminiCLIAvailable,
+        configure: registerGeminiCLI,
+        notInstalledHint: 'Not installed (https://github.com/google-gemini/gemini-cli)',
+    },
     {
         name: 'Codex',
         id: 'codex',
+        defaultSelected: false,
         detect: isCodexAvailable,
         configure: (token) => configureTomlMcp(join(homedir(), '.codex', 'config.toml'), token),
         notInstalledHint: 'Not installed (https://openai.com/codex)',
@@ -838,7 +894,7 @@ export async function runSetup() {
     console.log('');
     const pickerItems = detected.map((d) => ({
         label: d.client.name,
-        selected: d.available, // pre-select detected clients
+        selected: d.available && (d.client.defaultSelected !== false),
         available: d.available,
         hint: d.available ? undefined : d.client.notInstalledHint,
     }));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@rvry/mcp",
-  "version": "0.4.2",
+  "version": "0.5.0",
   "description": "RVRY reasoning depth enforcement (RDE) engine client.",
   "type": "module",
   "bin": {