Skip to content
Back to skills

564 Mcp Setup 0f042270

DSecurity

How to configure Aleph as an MCP server in **all major MCP-compatible clients**: Cursor, VS Code, Claude Desktop, Codex CLI, Windsurf, Kimi CLI, and others. ---

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 11, 2026
ai-agentspythonphpshellbashgitapifrontendbackendsecuritydocumentation

Works with

  • claude desktop
  • cursor
  • vscode
  • terminal
  • cli
  • api
  • mcp

Security analysis

D42/100
  • criticalModifies startup scripts or system services for persistence
  • criticalModifies startup scripts or system services for persistence
  • mediumInstalls packages at runtime which could introduce malicious dependencies

Pro shows the line behind each finding and how to fix it

Scanned October 11, 2026

npx -y skills add tools-only/X-Skills --skill 564-mcp_setup_0f042270 --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of 564 Mcp Setup 0f042270?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for 564 Mcp Setup 0f042270
[![Security: D — Skills Directory](https://www.skillsdirectory.com/api/skills/tools-only-564-mcp-setup-0f042270/badge)](https://www.skillsdirectory.com/skills/tools-only-564-mcp-setup-0f042270)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

SKILL.md
# MCP Server Configuration Guide

How to configure Aleph as an MCP server in **all major MCP-compatible clients**:
Cursor, VS Code, Claude Desktop, Codex CLI, Windsurf, Kimi CLI, and others.

---

## Quick Start (Full Power Mode)

For maximum capability without needing to configure workspace roots:

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": ["--enable-actions", "--workspace-mode", "any", "--tool-docs", "concise"]
    }
  }
}
```

This enables:

- **All action tools** (`read_file`, `write_file`, `run_command`, `run_tests`)
- **Any git repo access** (not limited to a single workspace root)
- **Concise tool descriptions** (cleaner MCP tool list)

---

## Shared Sub-Query Sessions (Live Sandbox)

If you want CLI sub-agents spawned via `sub_query` to access the **same live
Aleph session** (tools, contexts, and sandbox state), enable streamable HTTP
sharing:

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": ["--enable-actions", "--workspace-mode", "any", "--tool-docs", "concise"],
      "env": {
        "ALEPH_SUB_QUERY_SHARE_SESSION": "true",
        "ALEPH_SUB_QUERY_HTTP_PORT": "8765",
        "ALEPH_SUB_QUERY_MCP_SERVER_NAME": "aleph_shared"
      }
    }
  }
}
```

Notes:

- The Aleph server spins up a **local streamable HTTP endpoint** on demand.
- CLI sub-agents (claude/codex/gemini) are pointed at that live server
  automatically.
- Customize host/path with `ALEPH_SUB_QUERY_HTTP_HOST` and
  `ALEPH_SUB_QUERY_HTTP_PATH` if needed.
- Tools are exposed under the server name you choose (default: `aleph_shared`).
- `aleph_shared` avoids conflicts with an existing `aleph` stdio entry in Codex
  config.
- For Claude, the shared session is passed via `--mcp-config` and
  `--strict-mcp-config` flags.

For even higher limits:

```json
{
  "args": [
    "--enable-actions", "--workspace-mode", "any", "--tool-docs", "concise",
    "--timeout", "120", "--max-output", "100000"
  ]
}
```

---

## Per-Client Configuration

Every client below uses the same core args. Only the file location and format
differ. Replace `/path/to/your-project` with your actual project root.

### Cursor

Config file:

- **macOS / Linux:** `~/.cursor/mcp.json`
- **Windows:** `%USERPROFILE%\.cursor\mcp.json`

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": [
        "--workspace-root", "/path/to/your-project",
        "--enable-actions",
        "--tool-docs", "concise"
      ]
    }
  }
}
```

### VS Code

Config file:

- **macOS / Linux:** `~/.vscode/mcp.json`
- **Windows:** `%USERPROFILE%\.vscode\mcp.json`

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": [
        "--workspace-root", "/path/to/your-project",
        "--enable-actions",
        "--tool-docs", "concise"
      ]
    }
  }
}
```

### Claude Desktop

Config file:

- **macOS / Linux:** `~/.claude/settings.json`
- **Windows:** `%USERPROFILE%\.claude\settings.json`

**Auto-discovery (simplest):**

```bash
aleph-rlm install claude-code
```

Then restart Claude Desktop.

**Manual:**

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": [
        "--workspace-root", "/path/to/your-project",
        "--enable-actions",
        "--tool-docs", "concise"
      ]
    }
  }
}
```

<details>
<summary><strong>Installing the Claude Desktop skill</strong></summary>

**Option 1:** Download [`docs/prompts/aleph.md`](docs/prompts/aleph.md) and
save to:

- macOS / Linux: `~/.claude/commands/aleph.md`
- Windows: `%USERPROFILE%\.claude\commands\aleph.md`

**Option 2:** From installed package:

```bash
# macOS / Linux
mkdir -p ~/.claude/commands
cp "$(python -c "import aleph; print(aleph.__path__[0])")/../docs/prompts/aleph.md" \
  ~/.claude/commands/aleph.md
```

```powershell
# Windows (PowerShell)
New-Item -ItemType Directory -Force -Path "$env:USERPROFILE\.claude\commands"
$alephPath = python -c "import aleph; print(aleph.__path__[0])"
Copy-Item "$alephPath\..\docs\prompts\aleph.md" "$env:USERPROFILE\.claude\commands\aleph.md"
```

This enables the `/aleph` command for structured reasoning workflows.

</details>

### Codex CLI

Config file:

- **macOS / Linux:** `~/.codex/config.toml`
- **Windows:** `%USERPROFILE%\.codex\config.toml`

```toml
[mcp_servers.aleph]
command = "aleph"
args = ["--enable-actions", "--tool-docs", "concise"]
```

With a fixed workspace root:

```toml
[mcp_servers.aleph]
command = "aleph"
args = [
  "--workspace-root", "/path/to/your-project",
  "--enable-actions",
  "--tool-docs", "concise"
]
```

<details>
<summary><strong>Installing the Codex skill</strong></summary>

**Option 1:** Download [`docs/prompts/aleph.md`](docs/prompts/aleph.md) and
save to:

- macOS / Linux: `~/.codex/skills/aleph/SKILL.md`
- Windows: `%USERPROFILE%\.codex\skills\aleph\SKILL.md`

**Option 2:** From installed package:

```bash
# macOS / Linux
mkdir -p ~/.codex/skills/aleph
cp "$(python -c "import aleph; print(aleph.__path__[0])")/../docs/prompts/aleph.md" \
  ~/.codex/skills/aleph/SKILL.md
```

```powershell
# Windows (PowerShell)
New-Item -ItemType Directory -Force -Path "$env:USERPROFILE\.codex\skills\aleph"
$alephPath = python -c "import aleph; print(aleph.__path__[0])"
Copy-Item "$alephPath\..\docs\prompts\aleph.md" "$env:USERPROFILE\.codex\skills\aleph\SKILL.md"
```

This enables the `$aleph` command in Codex.

</details>

### Kimi CLI

Add via the command line:

```bash
kimi mcp add --transport stdio aleph -- \
  aleph --enable-actions --tool-docs concise --workspace-root /path/to/your-project
```

Or edit `~/.kimi/mcp.json` directly:

```json
{
  "mcpServers": {
    "aleph": {
      "transport": "stdio",
      "command": "aleph",
      "args": [
        "--enable-actions", "--tool-docs", "concise",
        "--workspace-root", "/path/to/your-project"
      ]
    }
  }
}
```

<details>
<summary><strong>Installing the Kimi skill</strong></summary>

Kimi CLI searches for skills in these locations (in order):

1. `~/.config/agents/skills/`
2. `.agents/skills/` (project root)
3. `~/.agents/skills/`
4. `~/.kimi/skills/`
5. `~/.claude/skills/`
6. `~/.codex/skills/`
7. `.kimi/skills/` (project root)
8. `.claude/skills/` (project root)
9. `.codex/skills/` (project root)

**User-level (recommended):**

```bash
mkdir -p ~/.config/agents/skills/aleph
cp "$(python -c "import aleph; print(aleph.__path__[0])")/../docs/prompts/aleph.md" \
  ~/.config/agents/skills/aleph/SKILL.md
```

**Project-level:**

```bash
mkdir -p .agents/skills/aleph
cp "$(python -c "import aleph; print(aleph.__path__[0])")/../docs/prompts/aleph.md" \
  .agents/skills/aleph/SKILL.md
```

Override the search path with `--skills-dir`.

</details>

### Windsurf

Standard MCP configuration:

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": [
        "--workspace-root", "/path/to/your-project",
        "--enable-actions",
        "--tool-docs", "concise"
      ]
    }
  }
}
```

### Cline / Continue.dev

These clients support standard MCP configuration. Check their documentation for
exact file locations and format.

### Generic MCP Client

Key parameters:

| Parameter              | Value                                            |
|------------------------|--------------------------------------------------|
| Command                | `aleph`                                          |
| Required args          | `--workspace-root /path/to/project`              |
| Optional args          | `--enable-actions`, `--tool-docs concise`, `--require-confirmation`, `--timeout N` |

---

## Parameters Reference

| Flag                                 | Default          | Description                                         |
|--------------------------------------|------------------|-----------------------------------------------------|
| `--workspace-root <path>`            | auto-detect      | Root directory for file operations                   |
| `--workspace-mode <fixed\|git\|any>` | `fixed`          | Path scope: single dir, any git repo, or unrestricted|
| `--enable-actions`                   | off              | Enable action tools (read/write/run)                 |
| `--require-confirmation`             | off              | Require `confirm=true` on action calls               |
| `--tool-docs <concise\|full>`        | `concise`        | Tool description verbosity                           |
| `--timeout <seconds>`                | 60               | Sandbox execution timeout                            |
| `--max-output <chars>`               | 50,000           | Max output characters from commands                  |
| `--max-file-size <bytes>`            | 1,000,000,000    | Max file size for read operations (1 GB)             |
| `--max-write-bytes <bytes>`          | 100,000,000      | Max file size for write operations (100 MB)          |

---

## Sub-Query Backends

`sub_query` can use an API backend or a local CLI backend. When
`ALEPH_SUB_QUERY_BACKEND` is `auto` (default), Aleph chooses the first
available:

1. **codex CLI** -- if installed
2. **gemini CLI** -- if installed
3. **claude CLI** -- if installed (deprioritized in MCP/sandbox contexts)
4. **API** -- if API credentials are available (fallback)

### API Configuration

The API backend uses **OpenAI-compatible endpoints only**:

| Variable                  | Fallback          | Description                                   |
|---------------------------|-------------------|-----------------------------------------------|
| `ALEPH_SUB_QUERY_API_KEY` | `OPENAI_API_KEY` | API key                                       |
| `ALEPH_SUB_QUERY_URL`     | `OPENAI_BASE_URL`| Base URL (default: `https://api.openai.com/v1`)|
| `ALEPH_SUB_QUERY_MODEL`   | --                | Model name (**required**)                     |

### Quick Setup Examples

```bash
# OpenAI
export ALEPH_SUB_QUERY_API_KEY=sk-...
export ALEPH_SUB_QUERY_MODEL=your-model-name

# Groq (fast inference)
export ALEPH_SUB_QUERY_API_KEY=gsk_...
export ALEPH_SUB_QUERY_URL=https://api.groq.com/openai/v1
export ALEPH_SUB_QUERY_MODEL=llama-3.3-70b-versatile

# Local LLM (Ollama, LM Studio, etc.)
# Make sure your local server is running and the model is available.
export ALEPH_SUB_QUERY_API_KEY=ollama   # any non-empty value
export ALEPH_SUB_QUERY_URL=http://localhost:11434/v1
export ALEPH_SUB_QUERY_MODEL=llama3.2
```

### All Sub-Query Variables

| Variable                          | Description                                                    |
|-----------------------------------|----------------------------------------------------------------|
| `ALEPH_SUB_QUERY_BACKEND`        | Force backend: `api`, `claude`, `codex`, `gemini`              |
| `ALEPH_SUB_QUERY_API_KEY`        | API key (fallback: `OPENAI_API_KEY`)                           |
| `ALEPH_SUB_QUERY_URL`            | API base URL (fallback: `OPENAI_BASE_URL`)                     |
| `ALEPH_SUB_QUERY_MODEL`          | Model name (required for API backend)                          |

Use `ALEPH_SUB_QUERY_BACKEND` to pin a backend or bypass auto-detection;
otherwise leave it unset for the default `auto` selection order.

> **Note:** Some MCP clients don't reliably pass `env` vars from their config to
> the server process. If `sub_query` reports "API key not found" despite your
> client's MCP settings, add the exports to your shell profile:
>
> - **macOS / Linux:** `~/.zshrc` or `~/.bashrc`
> - **Windows:** System Environment Variables or `$PROFILE` in PowerShell
>
> Then restart your terminal/client.

See [docs/CONFIGURATION.md](docs/CONFIGURATION.md) for the full list.

---

## Workspace Root

The workspace root should be the directory containing your `.git` folder,
`pyproject.toml`, `package.json`, etc.

**Automatic detection:** if you don't set `--workspace-root`, Aleph will:

1. Use `ALEPH_WORKSPACE_ROOT` if set
2. Prefer `PWD` (falls back to `INIT_CWD`) when present
3. Check if `.git` exists in that directory
4. If not, search parent directories until finding `.git`
5. Use that directory as the workspace root

**Recommended:** always set `--workspace-root` explicitly to avoid ambiguity.
For multi-repo work, prefer `--workspace-mode git` and use absolute paths (or a
broad workspace root).

---

## Example Scenarios

### Python Project

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": [
        "--workspace-root", "/Users/yourname/projects/my-python-app",
        "--enable-actions",
        "--tool-docs", "concise"
      ]
    }
  }
}
```

### Monorepo (scoped to a subdirectory)

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": [
        "--workspace-root", "/Users/yourname/monorepo/packages/frontend",
        "--enable-actions",
        "--tool-docs", "concise"
      ]
    }
  }
}
```

### Any Git Repo

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": [
        "--enable-actions",
        "--tool-docs", "concise",
        "--workspace-mode", "git"
      ]
    }
  }
}
```

### Increased Limits

```json
{
  "mcpServers": {
    "aleph": {
      "command": "aleph",
      "args": [
        "--workspace-root", "/path/to/project",
        "--enable-actions",
        "--tool-docs", "concise",
        "--timeout", "60",
        "--max-output", "100000",
        "--max-file-size", "5000000000",
        "--max-write-bytes", "500000000"
      ]
    }
  }
}
```

---

## Security

### Actions Mode

When you enable `--enable-actions`, you grant Aleph permission to:

| Capability        | Default Limit |
|-------------------|---------------|
| **Read files**    | Up to 1 GB    |
| **Write files**   | Up to 100 MB  |
| **Run commands**  | 60 s timeout  |
| **Run tests**     | 60 s timeout  |

Use `--workspace-mode git` to limit access to git repos, or
`--workspace-mode any` to remove path restrictions.

### Confirmation Mode

Use `--require-confirmation` for safer operation. When enabled, all action tools
require `confirm=true` in the call:

```json
{
  "args": [
    "--workspace-root", "/path/to/project",
    "--enable-actions",
    "--tool-docs", "concise",
    "--require-confirmation"
  ]
}
```

---

## Troubleshooting

### "Path escapes workspace root"

**Cause:** workspace root not set or incorrect.

**Fix:** add `--workspace-root` with the correct path, or use
`--workspace-mode git` / `--workspace-mode any` for multi-repo access.

### "Actions are disabled"

**Cause:** `--enable-actions` flag not set.

**Fix:** add `--enable-actions` to the `args` array in your MCP config.

### MCP Server Not Starting

Check in order:

1. Aleph is installed: `pip install "aleph-rlm[mcp]"`
2. Entry point works: `aleph --help`
3. Python is in PATH: try the full path to `python3`
4. Workspace root path is correct
5. MCP client was restarted after config changes

### sub_query Timed Out

Increase the timeout and/or reduce context size:

```bash
# Environment variable
export ALEPH_SUB_QUERY_TIMEOUT=120

# CLI flag
aleph --sub-query-timeout 120
```

**Runtime (MCP tool):**

```python
mcp__aleph__configure(sub_query_timeout=120)
```

For very large context slices, chunk first (e.g., `chunk(100000)` +
`sub_query_batch`).

### sub_query Reports "API Key Not Found"

**Cause:** some MCP clients don't pass `env` vars reliably.

**Fix:** add credentials to your shell profile:

```bash
# macOS / Linux (~/.zshrc or ~/.bashrc)
export ALEPH_SUB_QUERY_API_KEY=sk-...
export ALEPH_SUB_QUERY_MODEL=your-model-name
```

```powershell
# Windows (PowerShell $PROFILE)
$env:ALEPH_SUB_QUERY_API_KEY = "sk-..."
$env:ALEPH_SUB_QUERY_MODEL = "your-model-name"
```

Then restart your terminal/MCP client.

---

## Related Documentation

| Document                                                | Description                        |
|---------------------------------------------------------|------------------------------------|
| [README.md](README.md)                                  | Project overview and installation  |
| [docs/CONFIGURATION.md](docs/CONFIGURATION.md)          | Full configuration reference       |
| [DEVELOPMENT.md](DEVELOPMENT.md)                        | Architecture and contributing      |

---

## Support

For MCP configuration issues, check:

1. Your MCP client documentation (Cursor, VS Code, Claude Desktop, Codex, etc.)
2. This configuration guide
3. [docs/CONFIGURATION.md](docs/CONFIGURATION.md)

For Aleph-specific bugs or feature requests, open an issue on
[GitHub](https://github.com/Hmbown/aleph).

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.