Git Agent Protocol

https://www.gitagent.sh/
Admin
Git Agent Protocol is an open, git-native, framework-agnostic standard for defining, versioning, and...
Tokens:4,034
Snippets:66
Trust Score:6.7
Update:5 months ago
Show doc for...
Tokens:
Raw
### GitAgent CLI Quick Start Commands

Source: https://www.gitagent.sh/

A sequence of commands to install the GitAgent CLI, scaffold a new agent, validate its configuration, and run it with different models or export it.

```bash
$ npm install -g gitagent # Install the CLI
$ gitagent init --template standard # Scaffold an agent
$ gitagent validate # Check it's valid
$ gitagent run -d ./my-agent # Run with Claude Code
$ gitagent run -a github -p "Review" # Run with GitHub Models
$ gitagent export --format openai # Export for OpenAI SDK
$ gitagent run -r <repo> -a claude # Run with Claude Code

```

--------------------------------

### Install AI Agent Skill

Source: https://www.gitagent.sh/

Installs a specified AI agent skill globally or locally. Skills are reusable capability modules.

```bash
$ gitagent skills install code-review --global
```

--------------------------------

### Run a Git Agent

Source: https://www.gitagent.sh/

Execute a Git Agent Protocol agent from a GitHub repository and launch it with a specific agent type. Ensure the '@open-gitagent/gitagent' package is installed.

```bash
$ npx @open-gitagent/gitagent@latest run -r https://github.com/shreyas-lyzr/architect -a claude
```

--------------------------------

### Agent Configuration (agent.yaml)

Source: https://www.gitagent.sh/

Defines the core configuration for an agent, including its name, version, description, model preferences, skills, tools, and runtime parameters. This file is essential for agent setup.

```yaml
spec_version: "0.1.0"
name: code-review-agent
version: 1.0.0
description: Automated code review agent
author: gitagent-examples
license: MIT
model:
  preferred: claude-sonnet-4-5-20250929
  fallback:
    - claude-haiku-4-5-20251001
  constraints:
    temperature: 0.2
    max_tokens: 4096
skills:
  - code-review
tools:
  - lint-check
  - complexity-analysis
runtime:
  max_turns: 20
  timeout: 120
```

--------------------------------

### Install Git-Based Dependencies for Agents

Source: https://www.gitagent.sh/

Resolves and installs git-based dependencies for an agent. Dependencies are shallow-cloned into specified mount paths at their designated versions.

```bash
$ gitagent install
```

--------------------------------

### List Discovered AI Agent Skills

Source: https://www.gitagent.sh/

Displays a list of all AI agent skills that have been discovered or installed.

```bash
$ gitagent skills list
```

--------------------------------

### Scaffold a New Agent Repository

Source: https://www.gitagent.sh/

Initializes a new agent repository using one of the provided templates: minimal, standard, or full. Templates vary in included files like skills, tools, compliance, hooks, and memory.

```bash
$ gitagent init --template <minimal|standard|full>
```

--------------------------------

### Initialize LLM Wiki Agent

Source: https://www.gitagent.sh/

Scaffold a new agent with the LLM Wiki template to create a persistent knowledge base. This pattern is inspired by Karpathy's LLM Wiki.

```bash
gitagent init --template llm-wiki
```

--------------------------------

### Run Agent with Specified Adapter and Prompt

Source: https://www.gitagent.sh/

Executes an agent using a specified adapter (e.g., claude, openai, crewai) and a user-provided prompt. Supports auto-detection of the adapter.

```bash
$ gitagent run -a <adapter> -p "prompt"
```

--------------------------------

### Create, Update, and Run Agent on Lyzr Studio

Source: https://www.gitagent.sh/

A convenience command to clone a repository, create an agent on Lyzr Studio, and initiate a chat session. It saves the agent ID for future use.

```bash
$ gitagent lyzr run -r <repo> -p "Hello"
```

--------------------------------

### Import Agent Configuration from Another Format

Source: https://www.gitagent.sh/

Converts existing agent configurations from specified formats (e.g., Claude, Cursor, CrewAI) into the gitagent format.

```bash
$ gitagent import --from <format> <path>
```

--------------------------------

### Display Agent Summary Information

Source: https://www.gitagent.sh/

Shows a summary of the agent's configuration, including its model, skills, tools, compliance status, and a preview of the SOUL.md file.

```bash
$ gitagent info
```

--------------------------------

### Export Agent Definition to System Prompt Format

Source: https://www.gitagent.sh/

Concatenates the agent definition into a single markdown system prompt for any LLM.

```bash
$ gitagent export -f system-prompt
```

--------------------------------

### Git Agent Repository Structure

Source: https://www.gitagent.sh/

Illustrates a typical file and directory structure for a Git Agent Protocol repository, including configuration, core files, skills, tools, knowledge, memory, hooks, skillflows, and compliance.

```bash
$ tree my-agent-repository/
my-agent-repository/
├── agent.yamlconfig
├── SOUL.mdcore
├── RULES.md
├── AGENTS.md
├── INSTRUCTIONS.md
├── scheduler.yml
├── skills/capability
│ └── code-review/
│ └── SKILL.md
├── tools/capability
│ └── search.yaml
├── knowledge/
│ └── index.yaml
├── memory/
│ ├── MEMORY.md
│ └── runtime/live
│ ├── dailylog.md
│ ├── key-decisions.md
│ └── context.md
├── hooks/
│ ├── hooks.yaml
│ ├── bootstrap.md
│ └── teardown.md
├── skillflows/automation
│ └── code-review-flow.yaml
└── compliance/governance
└── regulatory-map.yaml
```

--------------------------------

### Export Agent Definition to Nanobot Format

Source: https://www.gitagent.sh/

Generates a configuration JSON and system prompt suitable for the Nanobot runtime.

```bash
$ gitagent export -f nanobot
```

--------------------------------

### SKILL.md Format Definition

Source: https://www.gitagent.sh/

Defines the structure and frontmatter for a SKILL.md file, including metadata like name, description, license, compatibility, author, version, and category, followed by instructional content.

```yaml
---
name: code-review
description: Thorough code reviews
license: MIT
compatibility: ">=0.1.0"
metadata:
  author: "Jane Doe"
  version: "1.0.0"
  category: "developer-tools"
---

# Instructions

Review the code for:
1. Security vulnerabilities
2. Performance issues
3. Code style consistency

```

--------------------------------

### Run AI Agent from Git Repository

Source: https://www.gitagent.sh/

Clones a specified agent repository and runs it with a given adapter and prompt. Supports specifying a branch, forcing a re-clone, or using a one-shot prompt.

```bash
gitagent run -r "https://github.com/shreyaskapale/shreyas-agent" -a claude
```

--------------------------------

### Search for AI Agent Skills

Source: https://www.gitagent.sh/

Searches for reusable capability modules (skills) within the SkillsMP marketplace, GitHub repositories, or the local filesystem using a keyword query.

```bash
$ gitagent skills search "code review"
```

--------------------------------

### Export Agent to Specified Format and Output Path

Source: https://www.gitagent.sh/

Exports the agent definition to a specified format (e.g., system-prompt, claude-code, openai) and saves it to an output file.

```bash
$ gitagent export --format <format> -o output
```

--------------------------------

### Define a SkillsFlow Workflow

Source: https://www.gitagent.sh/

This YAML defines a multi-step code review workflow. It includes steps for linting, agent-based review, testing, and reporting, with defined dependencies and error handling.

```yaml
name: code-review-flow
description: Full code review pipeline
triggers:
  - pull_request

steps:
  lint:
    skill: static-analysis
    inputs:
      path: ${{ trigger.changed_files }}

  review:
    agent: code-reviewer
    depends_on: [lint]
    prompt: |
      Focus on security and performance.
      Flag any use of eval() or raw SQL.
    inputs:
      findings: ${{ steps.lint.outputs.issues }}

  test:
    tool: bash
    depends_on: [lint]
    inputs:
      command: "npm test -- --coverage"

  report:
    skill: review-summary
    depends_on: [review, test]
    conditions:
      - ${{ steps.review.outputs.severity != 'none' }}
    inputs:
      review: ${{ steps.review.outputs.comments }}
      coverage: ${{ steps.test.outputs.report }}

error_handling:
  on_failure: notify
  channel: "#eng-reviews"

```

--------------------------------

### Inspect AI Agent Skill Metadata

Source: https://www.gitagent.sh/

Retrieves and displays the metadata for a specific AI agent skill, including its name, description, license, compatibility, and other details.

```bash
$ gitagent skills info code-review
```

--------------------------------

### Export Agent Definition to OpenClaw Format

Source: https://www.gitagent.sh/

Creates an OpenClaw workspace with configuration JSON, AGENTS.md, tools, and skills.

```bash
$ gitagent export -f openclaw
```

--------------------------------

### Export Agent Definition to Lyzr Studio Format

Source: https://www.gitagent.sh/

Creates an API payload for Lyzr Studio, including provider mappings and credential IDs.

```bash
$ gitagent export -f lyzr
```

--------------------------------

### Validate Agent Configuration Against Specification

Source: https://www.gitagent.sh/

Validates an agent's configuration against the Git Agent specification. Includes JSON schema validation, skill checks, and an optional regulatory compliance audit.

```bash
$ gitagent validate --compliance
```

--------------------------------

### Export Agent Definition to GitHub Models Format

Source: https://www.gitagent.sh/

Generates a chat completions payload for GitHub Models, including model namespace mappings.

```bash
$ gitagent export -f github
```

--------------------------------

### Export Agent Definition to OpenAI Agents SDK Format

Source: https://www.gitagent.sh/

Generates Python code for the OpenAI Agents SDK, including Agent() objects, Tool stubs, and type mappings.

```bash
$ gitagent export -f openai
```

--------------------------------

### Validate Agent Configuration

Source: https://www.gitagent.sh/

Run validation checks on the agent's configuration and behavior, typically used in CI/CD pipelines to ensure quality and catch violations before deployment.

```bash
gitagent validate
```

--------------------------------

### Export Agent Definition to Claude Code Format

Source: https://www.gitagent.sh/

Exports the agent definition, including skills, model hints, and compliance information, to the Claude Code format.

```bash
$ gitagent export -f claude-code
```

--------------------------------

### Generate Compliance Audit Report for Agent

Source: https://www.gitagent.sh/

Generates a compliance audit report for an AI agent, checking against standards such as FINRA 3110, SEC 17a-4, SR 11-7, and CFPB. The report includes pass/fail/warn indicators.

```bash
$ gitagent audit
```

--------------------------------

### Export Agent Definition to CrewAI Format

Source: https://www.gitagent.sh/

Generates a CrewAI YAML configuration, extracting roles, goals, and sub-agent mappings.

```bash
$ gitagent export -f crewai
```
Tokens:
Raw