# universal-tool-calling-protocol/code-mode

🔌 Plug-and-play library to enable agents to call MCP and UTCP tools via code execution.

Repository: https://github.com/universal-tool-calling-protocol/code-mode
Canonical: https://ross.abutalabs.com/products/code-mode
Language: TypeScript
License: MPL-2.0
License Family: copyleft
Topics: ai-agents, codemode, mcp, model-context-protocol, toolchain, utcp
Last push: 2026-08-25T16:39:29+00:00

## Health v2 (maintenance only)
Score: 69/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 58, longevity 21
- inputs: {"age_days": 295, "days_push": 8, "days_rel": 283, "gap_med": 7, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1544, forks 104 (observed 2026-08-28T04:05:01.291715+00:00)

## What it is
A TypeScript library that lets AI agents call MCP and UTCP tools by generating and executing code instead of traditional JSON tool-call round trips. It exposes the entire registered toolkit to the agent as a single code-execution tool, reducing iterations and token usage.

## Use cases
- let agents call MCP tools via code execution
- reduce tool-calling iterations for multi-tool workflows
- batch multiple tool calls in a single LLM response
- integrate UTCP and MCP toolchains into an agent
- cut token costs when agents orchestrate many tools

## When to choose
- your agent juggles many tools and suffers from multi-turn tool-call overhead
- you want a plug-and-play way to expose MCP/UTCP tools as one code-execution interface
- you're building agents in TypeScript/Node.js

## When to avoid
- you need agents to run in sandboxed environments where arbitrary code execution is unacceptable
- your agent only calls one or two tools per task
- you need a Python-first stack

## Facets
- artifact type: library
- maturity: active
- function: agent-framework, mcp, sdk, developer-tools
- domain: large-language-models, developer-tools
- platform: cross-platform
- tags: code-execution, tool-calling, utcp, codemode, llm-agents, ai-agents, automation, nodejs, typescript

## Member repositories
- universal-tool-calling-protocol/code-mode (main) score 69

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:01.291715+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:30:39.334868+00:00, confidence not recorded.
  - readme: https://github.com/universal-tool-calling-protocol/code-mode (fetched 2026-08-28T04:05:01.291715+00:00, sha 6d34c1178d99)
- Data as of 2026-08-30T08:39:29.467469+00:00.