CanerBaba25

RivalCode — AI skill for Claude Code

AI community

Open source profiler, replay debugger and benchmark arena for AI coding agents.

How to install RivalCode

This entry records only its repository, not the path inside it, so there is no exact command to give. Open CanerBaba25/RivalCode and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What RivalCode does

Open source profiler, replay debugger and benchmark arena for AI coding agents. Compare Codex, Claude Code, Gemini CLI and OpenCode on the same repository with real tests, diffs, costs, timelines and deterministic scoring.

Alternatives in AI

Manifest — Real-time cost observability for OpenClaw agents — track tokens, costs, messages, and model usage 4.1k ★
Token Monitor — Local-first desktop widget for tracking token usage, costs, and limits across 32+ AI coding tools—including Cl 1.8k ★
Oil UI — Push AI UI design to its limits: explore distinct directions, compare them side by side, and refine against re 861 ★

README

RivalCode

**Benchmark AI coding agents on your own repository.**

Codex · Claude Code · Gemini CLI · OpenCode

Same task. Same code. Same verification. Real evidence.

[![CI](https://github.com/CanerBaba25/RivalCode/actions/workflows/ci.yml/badge.svg)](https://github.com/CanerBaba25/RivalCode/actions/workflows/ci.yml) [![License: MIT](https://img.shields.io/badge/license-MIT-7dd3c7)](LICENSE) [![Node.js](https://img.shields.io/badge/node-%E2%89%A522.18-84cc16)](https://nodejs.org/) [![Local first](https://img.shields.io/badge/local--first-no%20telemetry-38bdf8)](SECURITY.md)

[Quick start](#quick-start) · [Try the demo](#try-the-demo) · [Dashboard](#dashboard) · [Documentation](#documentation)

RivalCode is a local profiler, replay debugger and benchmark arena for coding agents. Run a task, inspect what the agent actually did, and verify the resulting code yourself. Compare agents from the same commit in separate Git worktrees, with the same setup, checks and time limit.

No account, cloud backend or RivalCode telemetry. Your history lives in local SQLite; the selected agent keeps using its own CLI, authentication and permissions.

![RivalCode workspace overview showing real recorded Demo Agent runs](docs/images/overview.png)

The media use the explicit local Demo Agent. It performs real edits and runs real tests; provider performance, model tokens and cost are not simulated.

[Watch the local demo](docs/images/demo.webm) · [Compare recorded runs](docs/images/comparison.png) · [Inspect the battle](docs/images/dashboard.png) · [Open the saved diff](docs/images/replay-diff.png)

Quick start

Requires **Node.js 22.18+**, Git, pnpm 10, and at least one supported coding-agent CLI configured under its own account.

git clone https://github.com/CanerBaba25/RivalCode.git
cd RivalCode
pnpm install --frozen-lockfile
pnpm build

# Install the built CLI locally; no npm release is required.
pnpm pack --pack-destination .rivalcode/pa