firepass-mcp 0.1.3

MCP server that wraps Kimi K2.5 Turbo (Fireworks AI) as an agentic coding assistant with a full tool loop.

firepass-mcp

MCP server that turns Kimi K2.5 Turbo into an agentic coding assistant. The model gets a tool loop — it can read/write files, run shell commands, and search code with ripgrep, ast-grep, jq, and glob — and iterates autonomously until the task is done.

Three tools exposed over MCP:

Tool	Capabilities	Use case
firepass_worker	read, write, edit, bash, ripgrep, ast-grep, jq, glob, tree	Coding, refactoring, bug fixes
firepass_researcher	read, ripgrep, ast-grep, jq, glob, tree (read-only)	Code analysis, architecture review
firepass_reviewer	read, ripgrep, ast-grep, jq, glob, tree (read-only)	Code review with structured output

Requirements

• Python 3.10+
• A Fireworks AI API key
• rg (ripgrep), sg (ast-grep), jq, tree on PATH for full tool coverage
• bash, ls (standard on POSIX systems)

Install

uvx firepass-mcp

Configuration

Set your API key:

export FIREWORKS_API_KEY="fw-..."

Codex CLI

Add the server with:

codex mcp add firepass --env FIREWORKS_API_KEY=fw-... -- uv run firepass-mcp

This writes a config like:

[mcp_servers.firepass]
command = "uv"
args = ["run", "firepass-mcp"]

[mcp_servers.firepass.env]
FIREWORKS_API_KEY = "fw-..."

Claude Code

Add the server with:

claude mcp add -e FIREWORKS_API_KEY=fw-... firepass -- uv run firepass-mcp

This writes a config like:

{
  "mcpServers": {
    "firepass": {
      "type": "stdio",
      "command": "uv",
      "args": ["run", "firepass-mcp"],
      "env": {
        "FIREWORKS_API_KEY": "fw-..."
      }
    }
  }
}

Claude Desktop / Generic MCP JSON

If your client reads MCP JSON directly, use:

{
  "mcpServers": {
    "firepass": {
      "command": "uvx",
      "args": ["firepass-mcp"],
      "env": {
        "FIREWORKS_API_KEY": "fw-..."
      }
    }
  }
}

Environment variables

Variable	Default	Description
FIREWORKS_API_KEY	(required)	Fireworks AI API key
FIREPASS_MODEL	accounts/fireworks/routers/kimi-k2p5-turbo	Model ID
FIREPASS_BASH_TIMEOUT	60	Shell command timeout (seconds)
FIREPASS_MAX_OUTPUT	50000	Max chars per tool result
FIREPASS_MAX_READ	100000	Max chars per file read

How it works

1. You call firepass_worker, firepass_researcher, or firepass_reviewer with a prompt and a required cwd
2. The server sends the prompt to Kimi K2.5 Turbo with function-calling enabled
3. The model explores the codebase, makes edits, runs tests, and iterates
4. When done, it calls done() with an executive summary
5. The summary (plus an activity log) is returned as the tool result

All roles get 60 iterations by default, configurable per call.

Security model

All file operations (read_file, write_file, edit_file, glob_find, ripgrep, ast_grep, jq, tree, list_dir) are sandboxed to the required cwd you provide. Paths are resolved and validated against the working directory before any I/O.

The researcher and reviewer are read-only — bash, write_file, and edit_file are blocked both at the API schema level (model never sees them) and at runtime (server rejects them even if hallucinated). Dangerous ripgrep flags (--pre, --replace, -z) are also blocked.

The worker has full access including bash. It is not sandboxed at the command level — treat it like giving shell access to a remote developer scoped to your project directory.

Limits:

• File writes capped at 1 MB per operation
• File reads capped at 100K characters
• Tool output capped at 50K characters
• Context budget of 200K characters (old tool results truncated when exceeded)
• Configurable iteration limits (default 60 for all roles)

License

MIT