OpenAI to GigaChat migration skill

migrate-to-gigachat-skill

A Cursor skill that migrates Python, JS or TS projects from OpenAI to GigaChat: detects the stack, picks an adapter, and verifies with tests it still starts

Skill

Medium risk

We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.

Why this level

  • It changes the model provider and project dependencies, requiring tests before and after
  • It spends paid GigaChat tokens during the live check via a real model call
All reasons and checks
Russian stack

asya02/migrate-to-gigachat-skill

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add skills/migrate-to-gigachat-skill

Detects the agents on your machine, checks the risk and pins the version.

Other ways to install

Run in a terminal in the project folder

npx skills add asya02/migrate-to-gigachat-skill --skill migrate-to-gigachat-skill -a claude-code -y

The skills tool installs the current version from the repository. Add the -g flag to use the skill in every project.

Run in a terminal in the project folder

npx skills add asya02/migrate-to-gigachat-skill --skill migrate-to-gigachat-skill -a cursor -y

The skills tool installs the current version from the repository. Add the -g flag to use the skill in every project.

Run in a terminal in the project folder

npx skills add asya02/migrate-to-gigachat-skill --skill migrate-to-gigachat-skill -a github-copilot -y

The skills tool installs the current version from the repository. Add the -g flag to use the skill in every project.

Run in a terminal in the project folder

npx skills add asya02/migrate-to-gigachat-skill --skill migrate-to-gigachat-skill -a codex -y

The skills tool installs the current version from the repository. Add the -g flag to use the skill in every project.

Run in a terminal in the project folder

npx skills add asya02/migrate-to-gigachat-skill --skill migrate-to-gigachat-skill -a gemini-cli -y

The skills tool installs the current version from the repository. Add the -g flag to use the skill in every project.

Run in a terminal in the project folder

npx skills add asya02/migrate-to-gigachat-skill --skill migrate-to-gigachat-skill -a cline -y

The skills tool installs the current version from the repository. Add the -g flag to use the skill in every project.

Run in a terminal in the project folder

npx skills add asya02/migrate-to-gigachat-skill --skill migrate-to-gigachat-skill -a roo -y

The skills tool installs the current version from the repository. Add the -g flag to use the skill in every project.

A fork of Roo Code, same .roo folders.

Run in a terminal in the project folder

npx skills add asya02/migrate-to-gigachat-skill --skill migrate-to-gigachat-skill -a opencode -y

The skills tool installs the current version from the repository. Add the -g flag to use the skill in every project.

You will need: Node.js

Checked against the repository on Sep 25, 2026, commit c1e1f73.

Text for your agent

Copy the repository contents into the target project's .cursor/skills/migrate-to-gigachat-skill, run scripts/detect_stack.py to determine the migration path, set GIGACHAT_CREDENTIALS and GIGACHAT_SCOPE as environment variables, then ask the agent: migrate this project from OpenAI to GigaChat.

Other ways from the author
mkdir -p .cursor/skills/migrate-to-gigachat-skill && cp -r /path/to/this/repo/* .cursor/skills/migrate-to-gigachat-skill/

Copy the skill into the target project.

This is third-party code. Review the repository files before installing.

What it does

The skill detects how a project uses OpenAI, directly via the SDK, through LangChain/LangGraph, or through an OpenAI-compatible endpoint, and picks the matching migration path: for Python with LangChain it installs the langchain-gigachat package, for Python without LangChain it wires up the gpt2giga proxy for minimal code changes, for JS/TS with LangChainJS the same native integration applies, and for everything else in JS/TS it uses the gigachat npm SDK. Before and after the migration the skill runs baseline and after checks: tests, lint and a smoke run, to confirm the project actually starts rather than just compiling.

Who it is for. For developers who need to move a project from OpenAI to GigaChat with test-backed verification rather than doing it by hand and hoping.

Good fit when

  • The project uses the OpenAI SDK, LangChain, LangGraph or an OpenAI-compatible endpoint and needs to move to GigaChat
  • You want the migration backed by baseline and after test checks
  • You want a minimal-code-change path via the gpt2giga proxy

Not a fit when

  • Your project is not Python, JS or TS: the skill does not support other languages
  • You use an agent other than Cursor: the install targets .cursor/skills

Example request

Migrate this project from OpenAI to GigaChat and verify it still starts

Limitations

It only supports Python, JS and TS. GigaChat credentials (GIGACHAT_CREDENTIALS, GIGACHAT_SCOPE) are set only through local environment variables; the skill explicitly forbids pasting secrets into the chat. Strict mode (tests and lint must pass) is opt-in; verification is softer by default.

How to disable. Delete the .cursor/skills/migrate-to-gigachat-skill folder from the project.

Security check

  • It changes the model provider and project dependencies, requiring tests before and after
  • It spends paid GigaChat tokens during the live check via a real model call

README in short

The README gives a table of migration paths by stack and adapter, the project layout with detect_stack, collect_cred_requirements, run_verify and live_canary scripts, a step-by-step quick start from install to the after check, an environment variable table for run_verify.py, and an explicit security section forbidding secrets in chat.

SKILL.md

---
name: migrate-to-gigachat-skill
description: Migrates Python/JS/TS projects from OpenAI models/providers to GigaChat. Use when the user asks to replace OpenAI with GigaChat, migrate a LangChain/LangGraph app to GigaChat, or keep OpenAI-compatible clients by routing them through the gpt2giga proxy. Includes mandatory test and smoke-run verification to ensure the project still starts.
---

# Migrate OpenAI to GigaChat

## What this skill does

When a project currently uses OpenAI models/providers (direct OpenAI SDK, OpenAI-compatible endpoints, LangChain ChatOpenAI, LangGraph workflows, etc.), migrate it to GigaChat using the correct adapter:

- Python + LangChain / LangGraph: use langchain-gigachat.
- Python without LangChain/LangGraph (or any OpenAI-compatible client): use gpt2giga proxy for minimal code changes.
- JS/TS + LangChainJS/LangGraphJS: use langchain-gigachat from the ai-forever/langchainjs repo (npm package).
- JS/TS without LangChain: use gigachat SDK (npm package).

This skill always runs a baseline test, performs the migration, then re-runs tests and a smoke run to confirm the project still starts.

FAQ

Do I need to rewrite all the code by hand?

No, projects without LangChain can use the gpt2giga proxy with minimal changes; LangChain/LangGraph projects get a native adapter installed.

How do I give the agent GigaChat secrets?

You do not; the skill explicitly forbids pasting them into chat, only local environment variables are used.

Editors’ pick

Open-source personal AI assistant on your own machine: answers in Telegram, Slack, Discord and WhatsApp, extended with skills and plugins

CLIHigh risk390.7KRepository stars
Editors’ pick

A self-improving agent from Nous Research with a TUI, messaging gateway, cron jobs and skills it writes itself

CLIHigh risk249.8KRepository stars
Editors’ pick

An open source coding agent for the terminal and desktop with build and plan modes

CLIHigh risk210.6KRepository stars
Editors’ pick

Open prompt library with a Claude Code plugin, MCP server and CLI: search, fetch and improve prompts and skills from an agent

PluginMedium risk171.5KRepository stars
Foxx AIOpenAI to GigaChat migration skill

I am Foxx AI and I have already vetted this tool. Ask about install, setup or anything else, and I will keep it simple.