Agent skill

vibegit

Stars 187
Forks 21

Install this agent skill to your Project

npx add-skill https://github.com/memovai/memov/tree/main/skills/vibegit

SKILL.md

VibeGit - AI Coding History

AI-assisted version control using Memov. Track every AI interaction with your codebase.

Prerequisites

This skill requires the memov Python package which provides the mem CLI command.

Installation Options

Choose one of the following methods:

bash
# Option 1: Official install script (recommended)
curl -fsSL https://raw.githubusercontent.com/memovai/memov/main/install.sh | bash

Verify Installation

bash
mem

Initialize in Your Project

bash
cd your-project
mem init

How to Use

There are two ways to use VibeGit:

  1. Skill scripts (this folder): run ./scripts/*.sh to call the local mem CLI.
  2. MCP tools (Claude Code): use the MCP server tools like mcp__mem-mcp__snap.

They are related but not identical:

  • Skill scripts are a thin wrapper over mem CLI.
  • MCP tools may add extra behavior (e.g. auto-track untracked files, auto-sync for RAG features).

Commands

All commands should be run from the skill directory using the scripts provided.

Core Operations

Script Description Usage
init.sh Initialize memov in a project ./scripts/init.sh
track.sh Track new files ./scripts/track.sh file1.py file2.py -p "..." -r "..."
snap.sh Record a code change ./scripts/snap.sh --files "file1.py,file2.py" -p "What was asked" -r "What was done"
history.sh View AI coding history ./scripts/history.sh [--limit 20]
show.sh Show commit details ./scripts/show.sh <commit_hash>
status.sh Check working directory status ./scripts/status.sh

Navigation

Script Description Usage
jump.sh Jump to a specific snapshot ./scripts/jump.sh <commit_hash>
branch.sh List/create/delete branches ./scripts/branch.sh [name] [--delete name]
switch.sh Switch branches ./scripts/switch.sh <branch_name>

Web UI

Before running mem ui start, confirm where you want to open the UI:

  • Which project directory should the UI read from?
    • If not sure, run pwd and use --loc <that path>.

Suggested confirmation prompt:

Script Description Usage
ui_start.sh Start visual history browser ./scripts/ui_start.sh [--loc /path] [--port 38888] [--foreground]
ui_stop.sh Stop the web server ./scripts/ui_stop.sh [--loc /path]
ui_status.sh Check server status ./scripts/ui_status.sh [--loc /path]

Automatic Recording

After every AI coding session, record the interaction:

bash
./scripts/snap.sh \
  --files "api.py,tests/test_api.py" \
  --prompt "Add authentication endpoint" \
  --response "Added /login POST endpoint with JWT token generation"

Parameters

  • --files: Comma-separated list of files that were modified
  • --prompt or -p: The user's original request
  • --response or -r: Summary of what was done
  • --by-user or -u: Mark as human edit (vs AI)

RAG Features (Optional)

Some features (semantic search, validate, vibe_debug/vibe_search tools) require installing extra dependencies.

  • CLI: install with pip install memov[rag] then run mem sync once.
  • MCP: those tools may not appear unless RAG dependencies are installed.

Examples

Record a bug fix

bash
./scripts/snap.sh \
  --files "auth.py" \
  --prompt "Fix null pointer in login" \
  --response "Added null check for user object at L45"

View recent history

bash
./scripts/history.sh --limit 10

Jump to previous state

bash
./scripts/jump.sh a1b2c3d

Direct CLI Usage

You can also use the mem CLI directly:

bash
# Initialize
mem init

# Track new files
mem track file1.py file2.py -p "Initial tracking"

# Snapshot changes
mem snap --files file1.py -p "Added feature X" -r "Implemented..."

# View history
mem history

# Show specific commit
mem show a1b2c3d

# Jump to snapshot
mem jump a1b2c3d

What Gets Recorded

Each snapshot captures:

  • Prompt: What you asked the AI to do
  • Response: What the AI said it did
  • Files: Which files were changed
  • Diff: Actual code changes
  • Timestamp: When it happened
  • Source: AI or human

Benefits

  • Never lose context: Every AI interaction is recorded
  • Time travel: Jump to any point in your coding history
  • Understand changes: See not just what changed, but why

Troubleshooting

"memov CLI not found" error

If you see this error when running any script, memov is not installed. Run:

bash
# Quick check
./install.sh

# Or install manually
pip install memov

Scripts not executable

bash
chmod +x scripts/*.sh install.sh

Expand your agent's capabilities with these related and highly-rated skills.

davila7/claude-code-templates

verl-rl-training

Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.

23,776 2,298
Explore
davila7/claude-code-templates

openrlhf-training

High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.

23,776 2,298
Explore
davila7/claude-code-templates

gguf-quantization

GGUF format and llama.cpp quantization for efficient CPU/GPU inference. Use when deploying models on consumer hardware, Apple Silicon, or when needing flexible quantization from 2-8 bit without GPU requirements.

23,776 2,298
Explore
davila7/claude-code-templates

Claude Code Guide

Master guide for using Claude Code effectively. Includes configuration templates, prompting strategies "Thinking" keywords, debugging techniques, and best practices for interacting with the agent.

23,776 2,298
Explore
davila7/claude-code-templates

qdrant-vector-search

High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or scalable vector storage with Rust-powered performance.

23,776 2,298
Explore
davila7/claude-code-templates

behavioral-modes

AI operational modes (brainstorm, implement, debug, review, teach, ship, orchestrate). Use to adapt behavior based on task type.

23,776 2,298
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results