Tokenectomy-Labs/Tokenectomy

Overview

Tokenectomy is an open-source MCP server that autonomously processes AI agent error logs in machine-to-machine setups. It removes framework noise and redacts secrets before the logs reach the model's context window. The project is written in Rust and is available on GitHub.

Best for

Best for
Developers using AI agents over MCP who need to keep secrets out of context windows

Use cases

• Sanitize agent error logs before they are sent to an LLM
• Strip repetitive framework messages from MCP tool output
• Prevent API keys and tokens from appearing in model context

How to use

Install

npx -y tokenectomy-razor --mcp

Tools exposed

• get_error_context
• analyze_code
• apply_code_patch
• search_stack_overflow
• audit_context_health

Tested with

Claude Desktop, Cursor, Windsurf, Cline, VS Code, ChatGPT

Example client config

{\n  "mcpServers": {\n    "tokenectomy": {\n      "command": "npx",\n      "args": ["-y", "tokenectomy-razor", "--mcp"]\n    }\n  }\n}

Notes

Tokenectomy is an open-source MCP server that autonomously processes AI agent error logs in machine-to-machine setups. It removes framework noise and redacts secrets before the logs reach the model’s context window. The project is written in Rust and is available on GitHub.

4 stars on GitHub. Last updated 2026-09-15. Licensed MIT.

Use cases

• Sanitize agent error logs before they are sent to an LLM
• Strip repetitive framework messages from MCP tool output
• Prevent API keys and tokens from appearing in model context

Pros

• Automated log scrubbing between machines without manual effort
• Secret redaction reduces risk of credential leakage into prompts
• Rust implementation offers performance and memory safety

Cons

• Small project with only 4 stars and limited community support
• Focused solely on error logs and does not filter other context types
• Requires an MCP-compatible environment to operate

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.