@contextstream/mcp-server 0.4.22 → 0.4.23

package/LICENSE

@@ -1,21 +1,21 @@
-MIT License
-
-Copyright (c) 2025 ContextStream
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+MIT License
+
+Copyright (c) 2025 ContextStream
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -1,328 +1,330 @@
-# ContextStream MCP Server
-
-Persistent memory, semantic search, and code intelligence for any MCP-compatible AI tool (Cursor, Claude Code, Windsurf, VS Code, Claude Desktop, Codex CLI, etc.).
-
-ContextStream is a shared "brain" for AI-assisted development. It stores decisions, preferences, lessons, tasks, and project context — and lets your AI tools search and analyze your codebase with consistent context across sessions.
-
-**v0.4.x:** Consolidated domain tools (~11 tools) for **~75% lower tool-registry token overhead** vs previous versions.
-
----
-
-## Just Ask
-
-**You don't need to memorize tool names.** Describe what you want and your AI will call the right ContextStream tools:
-
-| You say… | ContextStream does… |
-|----------|---------------------|
-| "session summary" | Summarizes current workspace/project context |
-| "what did we decide about auth?" | Recalls decisions related to authentication |
-| "remember we're using PostgreSQL" | Captures that fact for future sessions |
-| "search for payment code" | Searches your codebase semantically/hybrid/keyword |
-| "what depends on UserService?" | Analyzes dependency graph & impact |
-
-> **Tip:** For best results, add the **recommended editor rules** so your AI reliably calls `session_init` / `context_smart` when appropriate:
-> https://contextstream.io/docs/quickstart
-
-![ContextStream in action](compare1.gif)
-
----
-
-## Choose Your Mode (Token Footprint)
-
-MCP clients often inject the tool catalog into the model context. v0.4.x is designed to keep that overhead small.
-
-| Mode | What it exposes | Best for | Enable |
-|------|-----------------|----------|--------|
-| **Consolidated** (default) | ~11 domain tools with `action` / `mode` dispatch | Most users (recommended) | `CONTEXTSTREAM_CONSOLIDATED=true` |
-| **Router** (extreme minimization) | ~2 meta-tools (`contextstream`, `contextstream_help`) | Tight context budgets / many MCP servers | `CONTEXTSTREAM_PROGRESSIVE_MODE=true` |
-| **Legacy** (granular tools) | Older `light/standard/complete` toolsets | Back-compat / old prompts | `CONTEXTSTREAM_CONSOLIDATED=false` |
-
-> **Note:** The env var name for Router mode is `CONTEXTSTREAM_PROGRESSIVE_MODE` (historical naming). It enables the ~2-tool "router" surface.
-
----
-
-## Features
-
-- **Consolidated domain tools** (v0.4.x): short tool list with action/mode dispatch
-- Session-aware context loading (`session_init`, `context_smart`)
-- Memory capture + recall (decisions, preferences, lessons, tasks, bugs)
-- Code search (semantic, hybrid, keyword, pattern)
-- Knowledge graph + code analysis (dependencies, impact, call paths, circular deps, unused code)
-- Graph ingestion for full graph builds (`graph(action="ingest")`)
-- Local repo ingestion for indexing (`project(action="ingest_local")`)
-- Auto-context: on first tool call in a new session, the server can auto-initialize context
-
----
-
-## Graph Tiers
-
-| Tier | Capabilities |
-|------|--------------|
-| **Pro (Graph-Lite)** | Module-level import graph, dependencies, and 1-hop impact |
-| **Elite/Team (Full Graph)** | Module + call + dataflow + type layers, plus full graph ingestion |
-
----
-
-## Requirements
-
-- **Node.js 18+**
-- A **ContextStream account**
-- Auth via **API key** or **JWT**
-
-Default API URL: `https://api.contextstream.io`
-
----
-
-## Quickstart (2 minutes)
-
-### 1) Run the Setup Wizard (recommended)
-
-The wizard:
-- Authenticates (browser/device login by default)
-- Creates/stores an API key
-- Installs recommended editor rules (optional)
-- Writes MCP config files for supported tools
-
-```bash
-npx -y @contextstream/mcp-server setup
-```
-
-**Useful flags:**
-
-| Flag | Description |
-|------|-------------|
-| `--dry-run` | Preview without writing files |
-
-**Notes:**
-- The wizard stores credentials at `~/.contextstream/credentials.json` for convenience. Delete it to force a fresh login.
-- Codex CLI MCP config is global-only (`~/.codex/config.toml`), so the wizard writes Codex config globally when selected.
-- Some tools still require UI/CLI setup (the wizard will tell you when it can't write a config).
-
-### 2) Run the MCP Server
-
-**Recommended** (works well with MCP configs):
-
-```bash
-npx -y @contextstream/mcp-server
-```
-
-**Or install globally:**
-
-```bash
-npm install -g @contextstream/mcp-server
-contextstream-mcp
-```
-
-### 3) Keeping Updated
-
-**If you use `npx`:** Restart your AI tool/editor and run ContextStream again
-(or pin the version: `npx -y @contextstream/mcp-server@0.4.3`)
-
-**If you installed globally:**
-
-```bash
-npm update -g @contextstream/mcp-server
-```
-
-After updating, restart your AI tool/editor so it reloads the tool catalog.
-
----
-
-## Configure Your MCP Client (Manual)
-
-> If you ran the setup wizard, you can usually skip this.
-
-### Cursor / Windsurf / Claude Desktop (JSON)
-
-These clients use an `mcpServers` JSON config:
-
-| Client | Config path |
-|--------|-------------|
-| **Cursor** | `~/.cursor/mcp.json` (global) or `.cursor/mcp.json` (project) |
-| **Windsurf** | `~/.codeium/windsurf/mcp_config.json` |
-| **Claude Desktop (macOS)** | `~/Library/Application Support/Claude/claude_desktop_config.json` |
-| **Claude Desktop (Windows)** | `%APPDATA%\Claude\claude_desktop_config.json` |
-
-**Consolidated (default):**
-
-```json
-{
-  "mcpServers": {
-    "contextstream": {
-      "command": "npx",
-      "args": ["-y", "@contextstream/mcp-server"],
-      "env": {
-        "CONTEXTSTREAM_API_URL": "https://api.contextstream.io",
-        "CONTEXTSTREAM_API_KEY": "YOUR_API_KEY"
-      }
-    }
-  }
-}
-```
-
-**Router mode (~2 meta-tools):**
-
-```json
-{
-  "mcpServers": {
-    "contextstream": {
-      "command": "npx",
-      "args": ["-y", "@contextstream/mcp-server"],
-      "env": {
-        "CONTEXTSTREAM_API_URL": "https://api.contextstream.io",
-        "CONTEXTSTREAM_API_KEY": "YOUR_API_KEY",
-        "CONTEXTSTREAM_PROGRESSIVE_MODE": "true"
-      }
-    }
-  }
-}
-```
-
-### VS Code (`.vscode/mcp.json`)
-
-VS Code uses a top-level `servers` map:
-
-```json
-{
-  "servers": {
-    "contextstream": {
-      "type": "stdio",
-      "command": "npx",
-      "args": ["-y", "@contextstream/mcp-server"],
-      "env": {
-        "CONTEXTSTREAM_API_URL": "https://api.contextstream.io",
-        "CONTEXTSTREAM_API_KEY": "YOUR_API_KEY"
-      }
-    }
-  }
-}
-```
-
-**Strong recommendation:** Use `inputs` so you don't commit secrets:
-
-```json
-{
-  "servers": {
-    "contextstream": {
-      "type": "stdio",
-      "command": "npx",
-      "args": ["-y", "@contextstream/mcp-server"],
-      "env": {
-        "CONTEXTSTREAM_API_URL": "https://api.contextstream.io",
-        "CONTEXTSTREAM_API_KEY": "${input:contextstreamApiKey}"
-      }
-    }
-  },
-  "inputs": [
-    {
-      "id": "contextstreamApiKey",
-      "type": "promptString",
-      "description": "ContextStream API Key",
-      "password": true
-    }
-  ]
-}
-```
-
-### Claude Code (CLI)
-
-**User scope (all projects):**
-
-```bash
-claude mcp add --transport stdio contextstream --scope user \
-  --env CONTEXTSTREAM_API_URL=https://api.contextstream.io \
-  --env CONTEXTSTREAM_API_KEY=YOUR_KEY \
-  -- npx -y @contextstream/mcp-server
-```
-
-**Router mode:**
-
-```bash
-claude mcp add --transport stdio contextstream --scope user \
-  --env CONTEXTSTREAM_API_URL=https://api.contextstream.io \
-  --env CONTEXTSTREAM_API_KEY=YOUR_KEY \
-  --env CONTEXTSTREAM_PROGRESSIVE_MODE=true \
-  -- npx -y @contextstream/mcp-server
-```
-
-> **Windows caveat** (native Windows, not WSL): if `npx` isn't found, use `cmd /c npx -y @contextstream/mcp-server` after `--`.
-
-### Codex CLI (`~/.codex/config.toml`)
-
-```toml
-[mcp_servers.contextstream]
-command = "npx"
-args = ["-y", "@contextstream/mcp-server"]
-
-[mcp_servers.contextstream.env]
-CONTEXTSTREAM_API_URL = "https://api.contextstream.io"
-CONTEXTSTREAM_API_KEY = "YOUR_API_KEY"
-```
-
----
-
-## Tool Overview (v0.4.x Consolidated)
-
-In consolidated mode, you call **domain tools** with `action` / `mode`:
-
-### Core
-
-| Tool | Description |
-|------|-------------|
-| `session_init` | Initialize workspace/project context |
-| `context_smart` | Retrieve the best bounded context for the current message |
-
-### Domain Tools
-
-| Tool | Description |
-|------|-------------|
-| `search` | `mode=semantic\|hybrid\|keyword\|pattern` |
-| `session` | `action=capture\|recall\|remember\|get_lessons\|capture_lesson\|...` |
-| `memory` | Events + nodes CRUD, decisions, lessons, etc. |
-| `graph` | Dependencies, impact, call_path, ingest, etc. |
-| `project` | Indexing, ingest_local, stats, files, etc. |
-| `workspace` | List, get, associate, bootstrap |
-| `reminder` | List, create, snooze, complete, dismiss |
-| `integration` | `provider=slack\|github`, search, activity, etc. |
-| `help` | Tools, auth, version, editor_rules |
-
-### Examples
-
-```
+# ContextStream MCP Server
+
+Persistent memory, semantic search, and code intelligence for any MCP-compatible AI tool (Cursor, Claude Code, Windsurf, VS Code, Claude Desktop, Codex CLI, etc.).
+
+ContextStream is a shared "brain" for AI-assisted development. It stores decisions, preferences, lessons, tasks, and project context — and lets your AI tools search and analyze your codebase with consistent context across sessions.
+
+**v0.4.x:** Consolidated domain tools (~11 tools) for **~75% lower tool-registry token overhead** vs previous versions.
+
+---
+
+## Just Ask
+
+**You don't need to memorize tool names.** Describe what you want and your AI will call the right ContextStream tools:
+
+| You say… | ContextStream does… |
+|----------|---------------------|
+| "session summary" | Summarizes current workspace/project context |
+| "what did we decide about auth?" | Recalls decisions related to authentication |
+| "remember we're using PostgreSQL" | Captures that fact for future sessions |
+| "search for payment code" | Searches your codebase semantically/hybrid/keyword |
+| "what depends on UserService?" | Analyzes dependency graph & impact |
+
+> **Tip:** For best results, add the **recommended editor rules** so your AI reliably calls `session_init` / `context_smart` when appropriate:
+> https://contextstream.io/docs/quickstart
+
+![ContextStream in action](compare1.gif)
+
+---
+
+## Choose Your Mode (Token Footprint)
+
+MCP clients often inject the tool catalog into the model context. v0.4.x is designed to keep that overhead small.
+
+| Mode | What it exposes | Best for | Enable |
+|------|-----------------|----------|--------|
+| **Consolidated** (default) | ~11 domain tools with `action` / `mode` dispatch | Most users (recommended) | `CONTEXTSTREAM_CONSOLIDATED=true` |
+| **Router** (extreme minimization) | ~2 meta-tools (`contextstream`, `contextstream_help`) | Tight context budgets / many MCP servers | `CONTEXTSTREAM_PROGRESSIVE_MODE=true` |
+| **Legacy** (granular tools) | Older `light/standard/complete` toolsets | Back-compat / old prompts | `CONTEXTSTREAM_CONSOLIDATED=false` |
+
+> **Note:** The env var name for Router mode is `CONTEXTSTREAM_PROGRESSIVE_MODE` (historical naming). It enables the ~2-tool "router" surface.
+
+---
+
+## Features
+
+- **Consolidated domain tools** (v0.4.x): short tool list with action/mode dispatch
+- Session-aware context loading (`session_init`, `context_smart`)
+- Memory capture + recall (decisions, preferences, lessons, tasks, bugs)
+- Code search (semantic, hybrid, keyword, pattern)
+- Knowledge graph + code analysis (dependencies, impact, call paths, circular deps, unused code)
+- Graph ingestion for full graph builds (`graph(action="ingest")`)
+- Local repo ingestion for indexing (`project(action="ingest_local")`)
+- Auto-context: on first tool call in a new session, the server can auto-initialize context
+
+> **⚠️ Search-First Rule:** For best results, your AI should use ContextStream `search(mode="hybrid")` **before** local tools like Glob/Grep/Read. This is enforced via editor rules and `context_smart` responses.
+
+---
+
+## Graph Tiers
+
+| Tier | Capabilities |
+|------|--------------|
+| **Pro (Graph-Lite)** | Module-level import graph, dependencies, and 1-hop impact |
+| **Elite/Team (Full Graph)** | Module + call + dataflow + type layers, plus full graph ingestion |
+
+---
+
+## Requirements
+
+- **Node.js 18+**
+- A **ContextStream account**
+- Auth via **API key** or **JWT**
+
+Default API URL: `https://api.contextstream.io`
+
+---
+
+## Quickstart (2 minutes)
+
+### 1) Run the Setup Wizard (recommended)
+
+The wizard:
+- Authenticates (browser/device login by default)
+- Creates/stores an API key
+- Installs recommended editor rules (optional)
+- Writes MCP config files for supported tools
+
+```bash
+npx -y @contextstream/mcp-server setup
+```
+
+**Useful flags:**
+
+| Flag | Description |
+|------|-------------|
+| `--dry-run` | Preview without writing files |
+
+**Notes:**
+- The wizard stores credentials at `~/.contextstream/credentials.json` for convenience. Delete it to force a fresh login.
+- Codex CLI MCP config is global-only (`~/.codex/config.toml`), so the wizard writes Codex config globally when selected.
+- Some tools still require UI/CLI setup (the wizard will tell you when it can't write a config).
+
+### 2) Run the MCP Server
+
+**Recommended** (works well with MCP configs):
+
+```bash
+npx -y @contextstream/mcp-server
+```
+
+**Or install globally:**
+
+```bash
+npm install -g @contextstream/mcp-server
+contextstream-mcp
+```
+
+### 3) Keeping Updated
+
+**If you use `npx`:** Restart your AI tool/editor and run ContextStream again
+(or pin the version: `npx -y @contextstream/mcp-server@0.4.3`)
+
+**If you installed globally:**
+
+```bash
+npm update -g @contextstream/mcp-server
+```
+
+After updating, restart your AI tool/editor so it reloads the tool catalog.
+
+---
+
+## Configure Your MCP Client (Manual)
+
+> If you ran the setup wizard, you can usually skip this.
+
+### Cursor / Windsurf / Claude Desktop (JSON)
+
+These clients use an `mcpServers` JSON config:
+
+| Client | Config path |
+|--------|-------------|
+| **Cursor** | `~/.cursor/mcp.json` (global) or `.cursor/mcp.json` (project) |
+| **Windsurf** | `~/.codeium/windsurf/mcp_config.json` |
+| **Claude Desktop (macOS)** | `~/Library/Application Support/Claude/claude_desktop_config.json` |
+| **Claude Desktop (Windows)** | `%APPDATA%\Claude\claude_desktop_config.json` |
+
+**Consolidated (default):**
+
+```json
+{
+  "mcpServers": {
+    "contextstream": {
+      "command": "npx",
+      "args": ["-y", "@contextstream/mcp-server"],
+      "env": {
+        "CONTEXTSTREAM_API_URL": "https://api.contextstream.io",
+        "CONTEXTSTREAM_API_KEY": "YOUR_API_KEY"
+      }
+    }
+  }
+}
+```
+
+**Router mode (~2 meta-tools):**
+
+```json
+{
+  "mcpServers": {
+    "contextstream": {
+      "command": "npx",
+      "args": ["-y", "@contextstream/mcp-server"],
+      "env": {
+        "CONTEXTSTREAM_API_URL": "https://api.contextstream.io",
+        "CONTEXTSTREAM_API_KEY": "YOUR_API_KEY",
+        "CONTEXTSTREAM_PROGRESSIVE_MODE": "true"
+      }
+    }
+  }
+}
+```
+
+### VS Code (`.vscode/mcp.json`)
+
+VS Code uses a top-level `servers` map:
+
+```json
+{
+  "servers": {
+    "contextstream": {
+      "type": "stdio",
+      "command": "npx",
+      "args": ["-y", "@contextstream/mcp-server"],
+      "env": {
+        "CONTEXTSTREAM_API_URL": "https://api.contextstream.io",
+        "CONTEXTSTREAM_API_KEY": "YOUR_API_KEY"
+      }
+    }
+  }
+}
+```
+
+**Strong recommendation:** Use `inputs` so you don't commit secrets:
+
+```json
+{
+  "servers": {
+    "contextstream": {
+      "type": "stdio",
+      "command": "npx",
+      "args": ["-y", "@contextstream/mcp-server"],
+      "env": {
+        "CONTEXTSTREAM_API_URL": "https://api.contextstream.io",
+        "CONTEXTSTREAM_API_KEY": "${input:contextstreamApiKey}"
+      }
+    }
+  },
+  "inputs": [
+    {
+      "id": "contextstreamApiKey",
+      "type": "promptString",
+      "description": "ContextStream API Key",
+      "password": true
+    }
+  ]
+}
+```
+
+### Claude Code (CLI)
+
+**User scope (all projects):**
+
+```bash
+claude mcp add --transport stdio contextstream --scope user \
+  --env CONTEXTSTREAM_API_URL=https://api.contextstream.io \
+  --env CONTEXTSTREAM_API_KEY=YOUR_KEY \
+  -- npx -y @contextstream/mcp-server
+```
+
+**Router mode:**
+
+```bash
+claude mcp add --transport stdio contextstream --scope user \
+  --env CONTEXTSTREAM_API_URL=https://api.contextstream.io \
+  --env CONTEXTSTREAM_API_KEY=YOUR_KEY \
+  --env CONTEXTSTREAM_PROGRESSIVE_MODE=true \
+  -- npx -y @contextstream/mcp-server
+```
+
+> **Windows caveat** (native Windows, not WSL): if `npx` isn't found, use `cmd /c npx -y @contextstream/mcp-server` after `--`.
+
+### Codex CLI (`~/.codex/config.toml`)
+
+```toml
+[mcp_servers.contextstream]
+command = "npx"
+args = ["-y", "@contextstream/mcp-server"]
+
+[mcp_servers.contextstream.env]
+CONTEXTSTREAM_API_URL = "https://api.contextstream.io"
+CONTEXTSTREAM_API_KEY = "YOUR_API_KEY"
+```
+
+---
+
+## Tool Overview (v0.4.x Consolidated)
+
+In consolidated mode, you call **domain tools** with `action` / `mode`:
+
+### Core
+
+| Tool | Description |
+|------|-------------|
+| `session_init` | Initialize workspace/project context |
+| `context_smart` | Retrieve the best bounded context for the current message |
+
+### Domain Tools
+
+| Tool | Description |
+|------|-------------|
+| `search` | `mode=semantic\|hybrid\|keyword\|pattern` |
+| `session` | `action=capture\|recall\|remember\|get_lessons\|capture_lesson\|...` |
+| `memory` | Events + nodes CRUD, decisions, lessons, etc. |
+| `graph` | Dependencies, impact, call_path, ingest, etc. |
+| `project` | Indexing, ingest_local, stats, files, etc. |
+| `workspace` | List, get, associate, bootstrap |
+| `reminder` | List, create, snooze, complete, dismiss |
+| `integration` | `provider=slack\|github`, search, activity, etc. |
+| `help` | Tools, auth, version, editor_rules |
+
+### Examples
+
+```
 search(mode="semantic", query="auth middleware", limit=3)
-memory(action="create_node", node_type="decision", title="Auth strategy", content="...")
-graph(action="impact", target="UserService")
-```
-
-**Full tool catalog:** https://contextstream.io/docs/mcp/tools
-
----
-
-## Authentication
-
-Set **one** of:
-
-| Variable | Use case |
-|----------|----------|
-| `CONTEXTSTREAM_API_KEY` | Recommended for local/dev |
-| `CONTEXTSTREAM_JWT` | Useful for hosted/user-session flows |
-
----
-
-## Environment Variables
-
-### Required
-
-| Variable | Description |
-|----------|-------------|
-| `CONTEXTSTREAM_API_URL` | Base API URL (default `https://api.contextstream.io`) |
-| `CONTEXTSTREAM_API_KEY` | API key (unless using JWT) |
-| `CONTEXTSTREAM_JWT` | JWT (unless using API key) |
-
-### Token + Tool Surface Controls
-
-| Variable | Description |
-|----------|-------------|
+memory(action="create_node", node_type="decision", title="Auth strategy", content="...")
+graph(action="impact", target="UserService")
+```
+
+**Full tool catalog:** https://contextstream.io/docs/mcp/tools
+
+---
+
+## Authentication
+
+Set **one** of:
+
+| Variable | Use case |
+|----------|----------|
+| `CONTEXTSTREAM_API_KEY` | Recommended for local/dev |
+| `CONTEXTSTREAM_JWT` | Useful for hosted/user-session flows |
+
+---
+
+## Environment Variables
+
+### Required
+
+| Variable | Description |
+|----------|-------------|
+| `CONTEXTSTREAM_API_URL` | Base API URL (default `https://api.contextstream.io`) |
+| `CONTEXTSTREAM_API_KEY` | API key (unless using JWT) |
+| `CONTEXTSTREAM_JWT` | JWT (unless using API key) |
+
+### Token + Tool Surface Controls
+
+| Variable | Description |
+|----------|-------------|
 | `CONTEXTSTREAM_CONSOLIDATED` | `true` (default in v0.4.x) uses consolidated domain tools |
 | `CONTEXTSTREAM_PROGRESSIVE_MODE` | Enables Router mode (~2 meta-tools) |
 | `CONTEXTSTREAM_CONTEXT_PACK` | Enable Context Pack for `context_smart` (code + graph + distill). Defaults to `true` |
@@ -332,74 +334,74 @@ Set **one** of:
 | `CONTEXTSTREAM_OUTPUT_FORMAT` | Output formatting; e.g., `compact` / `pretty` |
 | `CONTEXTSTREAM_SEARCH_LIMIT` | Default MCP search limit (default: 3) |
 | `CONTEXTSTREAM_SEARCH_MAX_CHARS` | Max chars per search result content (default: 400) |
-
-### Optional Defaults
-
-| Variable | Description |
-|----------|-------------|
-| `CONTEXTSTREAM_WORKSPACE_ID` | Default workspace fallback |
-| `CONTEXTSTREAM_PROJECT_ID` | Default project ID fallback |
-| `CONTEXTSTREAM_USER_AGENT` | Custom user agent string |
-| `CONTEXTSTREAM_PRO_TOOLS` | Comma-separated tool names treated as PRO |
-| `CONTEXTSTREAM_UPGRADE_URL` | Upgrade link for Free users calling PRO tools |
-
----
-
-## Migration Notes (pre-0.4.x → 0.4.x)
-
-Most workflows **just work**, but tool names change in consolidated mode.
-
-| Before (granular) | After (consolidated) |
-|-------------------|----------------------|
-| `search_semantic(query="auth")` | `search(mode="semantic", query="auth")` |
-| `session_capture(...)` | `session(action="capture", ...)` |
-| `graph_dependencies(...)` | `graph(action="dependencies", ...)` |
-
-If you rely on granular tool names, you can temporarily set:
-
-```bash
-CONTEXTSTREAM_CONSOLIDATED=false
-```
-
----
-
-## Troubleshooting
-
-| Issue | Solution |
-|-------|----------|
-| **Tools not appearing** | Restart the client after editing MCP config; confirm Node 18+ is available in the client runtime |
-| **Unauthorized / 401** | Verify `CONTEXTSTREAM_API_URL` + `CONTEXTSTREAM_API_KEY` (or JWT) |
-| **Wrong workspace/project** | Run `session_init` and/or associate your folder with the correct workspace |
-| **Client warns about tool context size** | Use Router mode (`CONTEXTSTREAM_PROGRESSIVE_MODE=true`), or keep consolidated mode and reduce schema/output verbosity |
-
----
-
-## Development
-
-```bash
-git clone https://github.com/contextstream/mcp-server.git
-cd mcp-server
-npm install
-npm run dev
-npm run typecheck
-npm run build
-```
-
----
-
-## Links
-
-| Resource | URL |
-|----------|-----|
-| **Website** | https://contextstream.io |
-| **Docs** | https://contextstream.io/docs/mcp |
-| **Tool Catalog** | https://contextstream.io/docs/mcp/tools |
-| **Pricing** | https://contextstream.io/pricing |
-| **npm** | https://www.npmjs.com/package/@contextstream/mcp-server |
-| **GitHub** | https://github.com/contextstream/mcp-server |
-
----
-
-## License
-
-MIT
+
+### Optional Defaults
+
+| Variable | Description |
+|----------|-------------|
+| `CONTEXTSTREAM_WORKSPACE_ID` | Default workspace fallback |
+| `CONTEXTSTREAM_PROJECT_ID` | Default project ID fallback |
+| `CONTEXTSTREAM_USER_AGENT` | Custom user agent string |
+| `CONTEXTSTREAM_PRO_TOOLS` | Comma-separated tool names treated as PRO |
+| `CONTEXTSTREAM_UPGRADE_URL` | Upgrade link for Free users calling PRO tools |
+
+---
+
+## Migration Notes (pre-0.4.x → 0.4.x)
+
+Most workflows **just work**, but tool names change in consolidated mode.
+
+| Before (granular) | After (consolidated) |
+|-------------------|----------------------|
+| `search_semantic(query="auth")` | `search(mode="semantic", query="auth")` |
+| `session_capture(...)` | `session(action="capture", ...)` |
+| `graph_dependencies(...)` | `graph(action="dependencies", ...)` |
+
+If you rely on granular tool names, you can temporarily set:
+
+```bash
+CONTEXTSTREAM_CONSOLIDATED=false
+```
+
+---
+
+## Troubleshooting
+
+| Issue | Solution |
+|-------|----------|
+| **Tools not appearing** | Restart the client after editing MCP config; confirm Node 18+ is available in the client runtime |
+| **Unauthorized / 401** | Verify `CONTEXTSTREAM_API_URL` + `CONTEXTSTREAM_API_KEY` (or JWT) |
+| **Wrong workspace/project** | Run `session_init` and/or associate your folder with the correct workspace |
+| **Client warns about tool context size** | Use Router mode (`CONTEXTSTREAM_PROGRESSIVE_MODE=true`), or keep consolidated mode and reduce schema/output verbosity |
+
+---
+
+## Development
+
+```bash
+git clone https://github.com/contextstream/mcp-server.git
+cd mcp-server
+npm install
+npm run dev
+npm run typecheck
+npm run build
+```
+
+---
+
+## Links
+
+| Resource | URL |
+|----------|-----|
+| **Website** | https://contextstream.io |
+| **Docs** | https://contextstream.io/docs/mcp |
+| **Tool Catalog** | https://contextstream.io/docs/mcp/tools |
+| **Pricing** | https://contextstream.io/pricing |
+| **npm** | https://www.npmjs.com/package/@contextstream/mcp-server |
+| **GitHub** | https://github.com/contextstream/mcp-server |
+
+---
+
+## License
+
+MIT

package/dist/index.js

@@ -4905,6 +4905,13 @@ function normalizeGraphTier(value) {
 }
 var AI_PLAN_TIMEOUT_MS = 5e4;
 var AI_PLAN_RETRIES = 0;
+var INGEST_BENEFITS = [
+  "Enable semantic code search across your entire codebase",
+  "Get AI-powered code understanding and context for your questions",
+  "Unlock dependency analysis and impact assessment",
+  "Allow the AI assistant to find relevant code without manual file navigation",
+  "Build a searchable knowledge base of your codebase structure"
+];
 var ContextStreamClient = class {
   constructor(config) {
     this.config = config;
@@ -5450,6 +5457,72 @@ var ContextStreamClient = class {
     uuidSchema.parse(projectId);
     return request(this.config, `/projects/${projectId}/index/status`, { method: "GET" });
   }
+  /**
+   * Check if project ingestion is recommended and return a recommendation.
+   * This is used by session_init to inform the AI about the project's index status.
+   *
+   * @param projectId - Project UUID
+   * @param folderPath - Optional folder path for generating the ingest command
+   * @returns IngestRecommendation with status, reason, and benefits if recommended
+   */
+  async checkIngestRecommendation(projectId, folderPath) {
+    try {
+      const status = await this.projectIndexStatus(projectId);
+      const data = status.data || status;
+      const indexedFiles = data.indexed_files ?? 0;
+      const lastUpdated = data.last_updated;
+      const statusDetail = data.status_detail;
+      const isStale = lastUpdated ? Date.now() - new Date(lastUpdated).getTime() > 24 * 60 * 60 * 1e3 : false;
+      const isRecent = lastUpdated ? Date.now() - new Date(lastUpdated).getTime() < 60 * 60 * 1e3 : false;
+      const ingestCommand = folderPath ? `project(action="ingest_local", path="${folderPath}")` : `project(action="ingest_local", path="<your_project_path>")`;
+      if (indexedFiles === 0 || statusDetail === "no_files_indexed") {
+        return {
+          recommended: true,
+          status: "not_indexed",
+          indexed_files: 0,
+          reason: "No files have been indexed yet. Ingesting your codebase will enable semantic search and AI-powered code understanding.",
+          benefits: INGEST_BENEFITS,
+          command: ingestCommand
+        };
+      }
+      if (isStale) {
+        return {
+          recommended: true,
+          status: "stale",
+          indexed_files: indexedFiles,
+          last_indexed: lastUpdated,
+          reason: `Index is over 24 hours old (${indexedFiles} files). Re-ingesting will capture any recent changes.`,
+          benefits: ["Capture recent code changes", "Update dependency analysis", "Improve search accuracy"],
+          command: ingestCommand
+        };
+      }
+      if (isRecent) {
+        return {
+          recommended: false,
+          status: "recently_indexed",
+          indexed_files: indexedFiles,
+          last_indexed: lastUpdated,
+          reason: `Project was recently indexed (${indexedFiles} files). No action needed.`
+        };
+      }
+      return {
+        recommended: false,
+        status: "indexed",
+        indexed_files: indexedFiles,
+        last_indexed: lastUpdated,
+        reason: `Project is indexed (${indexedFiles} files). Consider re-indexing if you've made significant changes.`
+      };
+    } catch (error) {
+      const ingestCommand = folderPath ? `project(action="ingest_local", path="${folderPath}")` : `project(action="ingest_local", path="<your_project_path>")`;
+      return {
+        recommended: true,
+        status: "not_indexed",
+        reason: "Unable to determine index status. Ingesting will ensure your codebase is searchable.",
+        benefits: INGEST_BENEFITS,
+        command: ingestCommand
+      };
+    }
+  }
   /**
    * Ingest files for indexing
    * This uploads files to the API for indexing
@@ -5784,20 +5857,45 @@ var ContextStreamClient = class {
             });
           }
         }
-        if (projectId && (params.auto_index === void 0 || params.auto_index === true)) {
-          context.indexing_status = "started";
-          const projectIdCopy = projectId;
-          const rootPathCopy = rootPath;
-          (async () => {
+        if (projectId) {
+          const autoIndex = params.auto_index === void 0 || params.auto_index === true;
+          if (autoIndex) {
+            context.indexing_status = "started";
+            context.ingest_recommendation = {
+              recommended: false,
+              status: "auto_started",
+              reason: "Background ingestion started automatically. Your codebase will be searchable shortly."
+            };
+            const projectIdCopy = projectId;
+            const rootPathCopy = rootPath;
+            (async () => {
+              try {
+                for await (const batch of readAllFilesInBatches(rootPathCopy, { batchSize: 50 })) {
+                  await this.ingestFiles(projectIdCopy, batch);
+                }
+                console.error(`[ContextStream] Background indexing completed for ${rootPathCopy}`);
+              } catch (e) {
+                console.error(`[ContextStream] Background indexing failed:`, e);
+              }
+            })();
+          } else {
             try {
-              for await (const batch of readAllFilesInBatches(rootPathCopy, { batchSize: 50 })) {
-                await this.ingestFiles(projectIdCopy, batch);
+              const recommendation = await this.checkIngestRecommendation(projectId, rootPath);
+              context.ingest_recommendation = recommendation;
+              if (recommendation.recommended) {
+                console.error(`[ContextStream] Ingest recommended for ${rootPath}: ${recommendation.status}`);
               }
-              console.error(`[ContextStream] Background indexing completed for ${rootPathCopy}`);
             } catch (e) {
-              console.error(`[ContextStream] Background indexing failed:`, e);
+              console.error(`[ContextStream] Failed to check ingest recommendation:`, e);
+              context.ingest_recommendation = {
+                recommended: true,
+                status: "not_indexed",
+                reason: "Unable to determine index status. Consider ingesting to enable code search.",
+                benefits: INGEST_BENEFITS,
+                command: `project(action="ingest_local", path="${rootPath}")`
+              };
             }
-          })();
+          }
         }
       } catch (e) {
         context.project_error = String(e);
@@ -5877,6 +5975,22 @@ var ContextStreamClient = class {
         await this._fetchSessionContextFallback(context, workspaceId, projectId, params);
       }
     }
+    if (projectId && !context.ingest_recommendation) {
+      try {
+        const recommendation = await this.checkIngestRecommendation(projectId, rootPath);
+        context.ingest_recommendation = recommendation;
+        if (recommendation.recommended) {
+          console.error(`[ContextStream] Ingest recommended for existing project ${projectId}: ${recommendation.status}`);
+        }
+      } catch (e) {
+        console.error(`[ContextStream] Failed to check ingest recommendation for existing project:`, e);
+        context.ingest_recommendation = {
+          recommended: false,
+          status: "indexed",
+          reason: 'Unable to determine exact index status. Use project(action="index_status") to check.'
+        };
+      }
+    }
     return context;
   }
   /**
@@ -6717,8 +6831,11 @@ var ContextStreamClient = class {
     let charsUsed = 0;
     const maxChars = maxTokens * 4;
     let candidateContext;
+    const CS_FIRST_RULE = "\u26A0\uFE0F STOP: Before Glob/Grep/Read \u2192 search(mode=hybrid) FIRST. Local tools ONLY if 0 results.";
+    const CS_FIRST_RULE_MINIFIED = "R:CS-first|Idx:project.index_status->ingest|NoLocalScanUnlessCSempty";
     if (format === "minified") {
-      const parts = [];
+      const parts = [CS_FIRST_RULE_MINIFIED];
+      charsUsed = CS_FIRST_RULE_MINIFIED.length + 1;
       for (const item of items) {
         const entry = `${item.type}:${item.value}`;
         if (charsUsed + entry.length + 1 > maxChars) break;
@@ -6726,9 +6843,10 @@ var ContextStreamClient = class {
         charsUsed += entry.length + 1;
       }
       context = parts.join("|");
-      candidateContext = items.map((i) => `${i.type}:${i.value}`).join("|");
+      candidateContext = [CS_FIRST_RULE_MINIFIED, ...items.map((i) => `${i.type}:${i.value}`)].join("|");
     } else if (format === "structured") {
-      const grouped = {};
+      const grouped = { R: [CS_FIRST_RULE] };
+      charsUsed = CS_FIRST_RULE.length + 10;
       for (const item of items) {
         if (charsUsed > maxChars) break;
         if (!grouped[item.type]) grouped[item.type] = [];
@@ -6736,14 +6854,15 @@ var ContextStreamClient = class {
         charsUsed += item.value.length + 5;
       }
       context = JSON.stringify(grouped);
-      const candidateGrouped = {};
+      const candidateGrouped = { R: [CS_FIRST_RULE] };
       for (const item of items) {
         if (!candidateGrouped[item.type]) candidateGrouped[item.type] = [];
         candidateGrouped[item.type].push(item.value);
       }
       candidateContext = JSON.stringify(candidateGrouped);
     } else {
-      const lines = ["[CTX]"];
+      const lines = [CS_FIRST_RULE, "", "[CTX]"];
+      charsUsed = CS_FIRST_RULE.length + 10;
       for (const item of items) {
         const line = `${item.type}:${item.value}`;
         if (charsUsed + line.length + 1 > maxChars) break;
@@ -6752,7 +6871,7 @@ var ContextStreamClient = class {
       }
       lines.push("[/CTX]");
       context = lines.join("\n");
-      const candidateLines = ["[CTX]"];
+      const candidateLines = [CS_FIRST_RULE, "", "[CTX]"];
       for (const item of items) {
         candidateLines.push(`${item.type}:${item.value}`);
       }
@@ -6761,7 +6880,9 @@ var ContextStreamClient = class {
     }
     if (context.length === 0 && withDefaults.workspace_id) {
       const wsHint = items.find((i) => i.type === "W")?.value || withDefaults.workspace_id;
-      context = format === "minified" ? `W:${wsHint}|[NO_MATCHES]` : `[CTX]
+      context = format === "minified" ? `${CS_FIRST_RULE_MINIFIED}|W:${wsHint}|[NO_MATCHES]` : `${CS_FIRST_RULE}
+
+[CTX]
 W:${wsHint}
 [NO_MATCHES]
 [/CTX]`;
@@ -7619,16 +7740,18 @@ Only after this preflight, proceed with search/analysis below.
 
 ### Search & Code Intelligence (ContextStream-first)
 
+\u26A0\uFE0F **STOP: Before using Glob/Grep/Read/Explore** \u2192 Call \`search(mode="hybrid")\` FIRST. Use local tools ONLY if ContextStream returns 0 results.
+
 **Search order:**
 1. \`session(action="smart_search", query="...")\` - context-enriched
 2. \`search(mode="hybrid", query="...", limit=3)\` or \`search(mode="keyword", query="<filename>", limit=3)\`
 3. \`project(action="files")\` - file tree/list (only when needed)
 4. \`graph(action="dependencies", ...)\` - code structure
-5. Local repo scans (rg/ls/find) - only if ContextStream returns no results, errors, or the user explicitly asks
+5. Local repo scans (rg/ls/find) - ONLY if ContextStream returns no results, errors, or the user explicitly asks
 
 **Search defaults:** \`search\` returns the top 3 results with compact snippets. Use \`limit\` + \`offset\` for pagination, and \`content_max_chars\` to expand snippets when needed.
 
-If ContextStream returns results, stop and use them. Do NOT use local Search/Explore/Read unless you need exact code edits.
+If ContextStream returns results, stop and use them. NEVER use local Search/Explore/Read unless you need exact code edits or ContextStream returned 0 results.
 
 **Code Analysis:**
 - Dependencies: \`graph(action="dependencies", file_path="...")\`
@@ -7739,10 +7862,13 @@ Rules Version: ${RULES_VERSION}
 
 ### Behavior Rules
 
+\u26A0\uFE0F **STOP: Before using Glob/Grep/Read/Explore** \u2192 Call \`search(mode="hybrid")\` FIRST. Use local tools ONLY if ContextStream returns 0 results.
+
 - **First message**: Call \`session_init\` with context_hint, then call \`context_smart\` before any other tool or response
+- **On [INGEST_RECOMMENDED]**: Ask the user if they want to enable semantic code search. Explain: "Indexing your codebase enables AI-powered code search, dependency analysis, and better context. This takes a few minutes." If user agrees, run the provided \`project(action="ingest_local")\` command.
 - **Every message after**: Always call \`context_smart\` BEFORE responding (semantic search for relevant context)
 - **Before searching files/code**: Check \`project(action="index_status")\`; if missing/stale run \`project(action="ingest_local", path="<cwd>")\` or \`project(action="index")\`, and use \`graph(action="ingest")\` if needed
-- **For discovery**: Use \`session(action="smart_search")\` or \`search(mode="hybrid")\` before any local repo scans
+- **For discovery**: Use \`session(action="smart_search")\` or \`search(mode="hybrid")\` \u2014 NEVER use local Glob/Grep/Read first
 - **For file/function/config lookups**: Use \`search\`/\`graph\` first; only fall back to rg/ls/find if ContextStream returns no results
 - **If ContextStream returns results**: Do NOT use local Search/Explore/Read; only open specific files when needed for exact edits
 - **For code analysis**: Use \`graph(action="dependencies")\` or \`graph(action="impact")\` for call/dependency analysis
@@ -9743,6 +9869,9 @@ Access: ${accessLabel}${showUpgrade ? ` (upgrade: ${upgradeUrl2})` : ""}`;
         const isPlanLimit = String(errorCode).toUpperCase() === "FORBIDDEN" && String(errorMessage).toLowerCase().includes("plan limit reached");
         const upgradeHint = isPlanLimit ? `
 Upgrade: ${upgradeUrl2}` : "";
+        const isUnauthorized = String(errorCode).toUpperCase() === "UNAUTHORIZED";
+        const sessionHint = isUnauthorized ? `
+Hint: Run session_init(folder_path="<your_project_path>") first to establish a session, or check that your API key is valid.` : "";
         const errorPayload = {
           success: false,
           error: {
@@ -9751,7 +9880,7 @@ Upgrade: ${upgradeUrl2}` : "";
             details: errorDetails
           }
         };
-        const errorText = `[${errorCode}] ${errorMessage}${upgradeHint}${errorDetails ? `: ${JSON.stringify(errorDetails)}` : ""}`;
+        const errorText = `[${errorCode}] ${errorMessage}${upgradeHint}${sessionHint}${errorDetails ? `: ${JSON.stringify(errorDetails)}` : ""}`;
         return {
           content: [{ type: "text", text: errorText }],
           structuredContent: errorPayload,
@@ -11172,8 +11301,12 @@ Runs in the background and returns immediately; use 'projects_index_status' to m
       title: "Initialize conversation session",
       description: `Initialize a new conversation session and automatically retrieve relevant context.
 This is the FIRST tool AI assistants should call when starting a conversation.
-Returns: workspace info, project info, recent memory, recent decisions, relevant context, and high-priority lessons.
-Automatically detects the IDE workspace/project path and can auto-index code.
+Returns: workspace info, project info, recent memory, recent decisions, relevant context, high-priority lessons, and ingest_recommendation.
+
+The ingest_recommendation field indicates if the project needs indexing for code search:
+- If [INGEST_RECOMMENDED] appears, ask the user if they want to enable semantic code search
+- Benefits: AI-powered code understanding, dependency analysis, better context retrieval
+- If user agrees, run: project(action="ingest_local", path="<project_path>")
 
 IMPORTANT: Pass the user's FIRST MESSAGE as context_hint to get semantically relevant context!
 Example: session_init(folder_path="/path/to/project", context_hint="how do I implement auth?")
@@ -11306,10 +11439,24 @@ This does semantic search on the first message. You only need context_smart on s
       if (versionNotice?.behind) {
         noticeLines.push(`[VERSION_NOTICE] current=${versionNotice.current} latest=${versionNotice.latest} upgrade="${versionNotice.upgrade_command}"`);
       }
+      const ingestRec = result.ingest_recommendation;
+      if (ingestRec?.recommended) {
+        const benefitsList = ingestRec.benefits?.slice(0, 3).map((b) => `  \u2022 ${b}`).join("\n") || "";
+        noticeLines.push(
+          `[INGEST_RECOMMENDED] status=${ingestRec.status}`,
+          `Reason: ${ingestRec.reason}`,
+          ingestRec.benefits ? `Benefits:
+${benefitsList}` : "",
+          `Action: Ask the user if they want to enable code search by running:`,
+          `  ${ingestRec.command || 'project(action="ingest_local", path="<project_path>")'}`
+        );
+      } else if (ingestRec?.status === "auto_started") {
+        noticeLines.push(`[INGEST_STATUS] Background indexing started. Codebase will be searchable shortly.`);
+      }
       if (noticeLines.length > 0) {
         text = `${text}
 
-${noticeLines.join("\n")}`;
+${noticeLines.filter(Boolean).join("\n")}`;
       }
       return { content: [{ type: "text", text }], structuredContent: toStructured(result) };
     }

package/package.json

@@ -1,57 +1,57 @@
-{
-  "name": "@contextstream/mcp-server",
-  "mcpName": "io.github.contextstreamio/mcp-server",
-  "version": "0.4.22",
-  "description": "ContextStream MCP server - v0.4.x with consolidated domain tools (~11 tools, ~75% token reduction). Code context, memory, search, and AI tools.",
-  "type": "module",
-  "license": "MIT",
-  "main": "dist/index.js",
-  "files": [
-    "dist",
-    "README.md"
-  ],
-  "bin": {
-    "contextstream-mcp": "dist/index.js"
-  },
-  "scripts": {
-    "dev": "tsx src/index.ts",
-    "build": "esbuild src/index.ts --bundle --platform=node --target=node18 --outfile=dist/index.js --format=esm --external:@modelcontextprotocol/sdk --banner:js=\"#!/usr/bin/env node\" && esbuild src/test-server.ts --bundle --platform=node --target=node18 --outfile=dist/test-server.js --format=esm --external:@modelcontextprotocol/sdk",
-    "start": "node dist/index.js",
-    "test-server": "tsx src/test-server.ts",
-    "test-server:start": "node dist/test-server.js",
-    "typecheck": "tsc --noEmit"
-  },
-  "dependencies": {
-    "@modelcontextprotocol/sdk": "^1.25.1",
-    "zod": "^3.23.8"
-  },
-  "devDependencies": {
-    "@types/node": "^20.10.0",
-    "esbuild": "^0.27.0",
-    "tsx": "^4.15.4",
-    "typescript": "^5.6.3"
-  },
-  "engines": {
-    "node": ">=18"
-  },
-  "keywords": [
-    "mcp",
-    "model-context-protocol",
-    "contextstream",
-    "ai",
-    "code-context",
-    "memory",
-    "knowledge-graph"
-  ],
-  "repository": {
-    "type": "git",
-    "url": "git+https://github.com/contextstream/mcp-server.git"
-  },
-  "homepage": "https://contextstream.io/docs/mcp",
-  "bugs": {
-    "url": "https://github.com/contextstream/mcp-server/issues"
-  },
-  "overrides": {
-    "qs": "6.14.1"
-  }
-}
+{
+  "name": "@contextstream/mcp-server",
+  "mcpName": "io.github.contextstreamio/mcp-server",
+  "version": "0.4.23",
+  "description": "ContextStream MCP server - v0.4.x with consolidated domain tools (~11 tools, ~75% token reduction). Code context, memory, search, and AI tools.",
+  "type": "module",
+  "license": "MIT",
+  "main": "dist/index.js",
+  "files": [
+    "dist",
+    "README.md"
+  ],
+  "bin": {
+    "contextstream-mcp": "dist/index.js"
+  },
+  "scripts": {
+    "dev": "tsx src/index.ts",
+    "build": "esbuild src/index.ts --bundle --platform=node --target=node18 --outfile=dist/index.js --format=esm --external:@modelcontextprotocol/sdk --banner:js=\"#!/usr/bin/env node\" && esbuild src/test-server.ts --bundle --platform=node --target=node18 --outfile=dist/test-server.js --format=esm --external:@modelcontextprotocol/sdk",
+    "start": "node dist/index.js",
+    "test-server": "tsx src/test-server.ts",
+    "test-server:start": "node dist/test-server.js",
+    "typecheck": "tsc --noEmit"
+  },
+  "dependencies": {
+    "@modelcontextprotocol/sdk": "^1.25.1",
+    "zod": "^3.23.8"
+  },
+  "devDependencies": {
+    "@types/node": "^20.10.0",
+    "esbuild": "^0.27.0",
+    "tsx": "^4.15.4",
+    "typescript": "^5.6.3"
+  },
+  "engines": {
+    "node": ">=18"
+  },
+  "keywords": [
+    "mcp",
+    "model-context-protocol",
+    "contextstream",
+    "ai",
+    "code-context",
+    "memory",
+    "knowledge-graph"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/contextstream/mcp-server.git"
+  },
+  "homepage": "https://contextstream.io/docs/mcp",
+  "bugs": {
+    "url": "https://github.com/contextstream/mcp-server/issues"
+  },
+  "overrides": {
+    "qs": "6.14.1"
+  }
+}