GitAgent: An Open Standard for Portable AI Agents in Git Repos
What GitAgent Does
GitAgent addresses the problem of AI agent framework lock-in by creating a portable agent definition that works across multiple platforms. Instead of rewriting agents when switching frameworks, developers can define agents as files in a git repository that export to Claude Code, OpenAI Agents SDK, CrewAI, Google ADK, LangChain, and others.
Core Specification
The standard uses three required files:
agent.yaml- Configuration fileSOUL.md- Personality and instructionsSKILL.md- Capabilities definition
Git-Native Benefits
By being git-native, GitAgent provides several built-in advantages:
- Version control for agent behavior - roll back bad prompts like reverting commits
- Branching for environment promotion (dev → staging → main)
- Human-in-the-loop via PRs - agents can open branches for human review
- Audit trail via git blame and git diff
- Agent forking and remixing - fork public agents, customize, and PR improvements back
- CI/CD integration with GitAgent validate in GitHub Actions
CLI Usage
The command-line interface allows direct execution of agent repositories:
npx @open-gitagent/gitagent run -r https://github.com/user/agent -a claudeOptional Compliance Features
For organizations needing compliance controls, GitAgent offers:
- Risk tier management
- Regulatory mappings (FINRA, SEC, SR 11-7)
- Audit reports via
GitAgent audit
Availability
The specification is available at https://gitagent.sh with code on GitHub. The developers are seeking feedback on schema design and adapter priorities for future development.
📖 Read the full source: HN AI Agents
👀 See Also
DELIGHT: Local Orchestrator Uses Multiple ChatGPT Sessions as Coordinated Agents
DELIGHT is a local orchestrator that runs multiple hidden ChatGPT browser sessions simultaneously, coordinating them like a team of agents without requiring API keys or GPU resources. It connects to OpenClaw as an action layer to apply changes to real files and run tests.
Eä: A SIMD Compiler for Python Written in Rust
A developer built Eä, a compiler for SIMD kernels in ~12k lines of Rust that generates shared libraries and Python wrappers from .ea files, achieving 6.6× speedups over NumPy without ctypes or build systems.
80-line Python script uses Claude to auto-generate internal link suggestions, cuts linking time from 2 hours to 8 minutes
A Reddit user built an 80-line Python script that feeds an article draft and sitemap to Claude, returning relevant internal link targets with suggested anchor text — reducing manual linking time from 2 hours to 8 minutes per article.
Multi-Agent Debate Approach Improves LLM Reasoning Quality
A developer experimented with a multi-agent debate approach using CyrcloAI, where different AI agents take on roles like analyst, critic, and synthesizer to critique each other's responses before producing a final answer, resulting in more structured and deliberate outputs.