OpenAI Code AssistantBLOCK
OpenAI Code Assistant Model Context Protocol (MCP) Server
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
[](https://mseep.ai/app/arthurcolle-openai-mcp)
A powerful Python recreation of Claude Code with enhanced real-time visualization, cost management, and Model Context Protocol (MCP) server capabilities. This tool provides a natural language interface for software development tasks with support for multiple LLM providers.
Key Features
- Multi-Provider Support: Works with OpenAI, Anthropic, and other LLM providers
- Model Context Protocol Integration:
- Run as an MCP server for use with Claude Desktop and other clients
- Connect to any MCP server with the built-in MCP client
- Multi-agent synchronization for complex problem solving
- Real-Time Tool Visualization: See tool execution progress and results in real-time
- Cost Management: Track token usage and expenses with budget controls
- Comprehensive Tool Suite: File operations, search, command execution, and more
- Enhanced UI: Rich terminal interface with progress indicators and syntax highlighting
- Context Optimization: Smart conversation compaction and memory management
- Agent Coordination: Specialized agents with different roles can collaborate on tasks
Installation
- Clone this repository
- Install dependencies:
pip install -r requirements.txt
- Create a
.envfile with your API keys:
# Choose one or more providers OPENAI_API_KEY=your_openai_api_key_here ANTHROPIC_API_KEY=your_anthropic_api_key_here # Optional model selection OPENAI_MODEL=gpt-4o ANTHROPIC_MODEL=claude-3-opus-20240229
Usage
CLI Mode
Run the CLI with the default provider (determined from available API keys):
python claude.py chat
Specify a provider and model:
p
93e32586d7d3OBSERVED · 2026-10-08Exposed tools (18)
14 read · 3 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
Bash | read | Executes a given bash command in a persistent shell session. |
Edit | write | This is a tool for editing files. |
GetConfiguration | read | Get configuration for connecting Claude Desktop to this MCP server. |
GetServerMetrics | read | Get server metrics and statistics. |
GlobTool | read | Fast file pattern matching tool that works with any codebase size. |
GrepTool | read | Fast content search tool that works with any codebase size. |
JinaFactCheck | read | Fact check a statement using Jina.ai |
JinaReadURL | read | Read and summarize a webpage using Jina.ai |
JinaSearch | read | Search the web for information using Jina.ai |
LS | read | Lists files and directories in a given path. |
MCPGetContext | read | Get context from the MCP server using a prompt template |
MCPGetPrompt | read | Get details of a specific prompt template from the MCP server |
MCPListPrompts | read | List available prompt templates from the MCP server |
Replace | write | Write a file to the local filesystem. Overwrites the existing file if there is one. |
ResetServerMetrics | destructive | Reset server metrics tracking. |
View | read | Reads a file from the local filesystem. The file_path parameter must be an absolute path, not a relative path. |
Weather | read | Gets the current weather for a location |
tool_executor | write | # Create a tool call in the format expected by ToolExecutionManager |
Trust audit
BLOCKgrade D · trust 68/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (3 observation(s))
- Network
- declared (6 observation(s))
- Shell
- declared (2 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (13)
"SELECT ", "CREATE TABLE", "const ", "let ", "var ", "function(", "=>"]__init__.cpython-312.pyc
mcp_server.cpython-312.pyc
__init__.cpython-312.pyc
serve.cpython-312.pyc
__init__.cpython-312.pyc
CACHE_DIR = "/root/.cache"
RESULTS_DIR = "/root/results"
ResetServerMetrics
conversation_hash = hashlib.md5(serialized).hexdigest()
openai, anthropic, python-dotenv, pydantic, requests, fastmcp, typer, rich
Load environment variables from file:
Gates applied: no_behavioural_pass, no_license.
93e32586d7d3full audit observations/trust-audit/mcp-server/arthurcolle__openai-code-assistant.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-08 | 93e32586d7d3 | BLOCK | D | 68 | first audit |
Questions
What is the OpenAI Code Assistant MCP server?
OpenAI Code Assistant Model Context Protocol (MCP) Server
What tools does OpenAI Code Assistant expose?
18 in total: 14 read-only, 3 that write, and 1 that can delete or overwrite (ResetServerMetrics). Every one is listed on this page with its risk.
Is OpenAI Code Assistant safe to connect to an agent?
No — not without reading the findings first. The audit graded it D (68/100) and found 1 critical or high issue in the source. Each one is listed on this page with the file and line it is on. Separately from the audit: 1 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does OpenAI Code Assistant need?
It reads ANTHROPIC_API_KEY, BING_SEARCH_API_KEY, GOOGLE_SEARCH_API_KEY, JINA_API_KEY, MODAL_API_KEY, OPENAI_API_KEY and OPENWEATHER_API_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How current is this page?
The grade is for one exact copy of the source (93e32586d7d3), read on 2026-10-08. The repository is watched and re-audited when it changes.