😤 Why git-oops?

For Humans:

Your script crashes with a 50-line traceback. You're exhausted, and ChatGPT gives you a polite 6-paragraph email instead of a 2-line fix. git-oops roasts your sloppy code with razor-sharp wit and applies a surgical 1-click git patch directly in your terminal.

For All Autonomous AI Coding Agents (Antigravity, Claude, Cursor, Windsurf, Qoder, Codex, Devin, and any Agent):

Modern autonomous coding agents face catastrophic failure modes:

1. Context Bloat: Multi-thousand line terminal dumps overflow agent context windows.
2. Infinite Thrashing: Agents get trapped in loops editing and reverting the same code.
3. Ghost Hallucinations: Agents import non-existent libraries or fall into typosquatted package traps.
4. Blast-Radius Disasters: Agents accidentally execute rm -rf or destructive force-pushes.

git-oops provides the missing runtime safety harness that shields your repository, your sanity, and your wallet.

⚡ The Dual-Mode Architecture

                        [ Terminal Crash / Agent Execution ]
                                          │
                   ┌──────────────────────┴──────────────────────┐
                   ▼                                             ▼
        [ 🧑‍💻 HUMAN MODE ]                                [ 🤖 AGENT MODE ]
  • High-Fidelity Rounded Cards                   • Clean Machine-Readable JSON
  • Caustic Critic Roasts (Staff/Chef)            • 90% Context Log Squeezer & Redactor
  • Interactive [y/N] Patch Confirmation          • Blast-Radius Firewall & AST Linter
  • Instant `--undo` Reversion                    • Anti-Thrashing Watchdog & Mutex
                   │                                             │
                   └──────────────────────┬──────────────────────┘
                                          ▼
                         [ 🛡️ SHA-256 Cache & Circuit Breaker ]
                         [ 🔌 Native Model Context Protocol (MCP) ]

🌟 The 20 Super-Capabilities

🚀 Beginner Quickstart (1-Minute Guide)

1. Install Globally:

pip install git-oops

Zero Configuration Required: git-oops automatically registers machine-wide hooks in the background. Every git commit and git push across your repositories is immediately shielded with sub-second L3 security.

2. Live Obsidian Web HUD:

Launch the real-time glassmorphism dashboard in one command:

git-oops --hud

3. Connect to Any AI Tool (MCP or Subshell):

Option A: For Antigravity, Claude Desktop, Cursor, Windsurf, Qoder & All MCP Agents:

Add this to your MCP configuration:

{
  "mcpServers": {
    "git-oops": {
      "command": "git-oops",
      "args": ["--mcp"]
    }
  }
}

Option B: For Any Subshell / CLI AI Agent (Codex, Devin, Copilot, Aider, Roo Code, Cline, etc.):

Simply tell your agent in prompt or rule:

"Always diagnose and apply patches safely via: git-oops --agent --file <file> --error \"<error>\" --yes" It outputs clean machine-readable JSON without ANSI codes.

4. Interactive Terminal Commands:

# Human mode (roasts you & shows visual diff)
git-oops --file broken.py --error "KeyError: 'user_id'"

# 1ms Zero-AI Symbolic AST repair ($0.00 cost)
git-oops --file broken.py --error "AttributeError: 'NoneType' object has no attribute 'x'" --solver --yes

# Run full L3 security pre-flight manually
git-oops --pre-commit

# Pre-PR Staff Gatekeeper audit
git-oops --review-pr

# Resurrect lost work after an agent wipeout
git-oops --salvage

🧪 Comprehensive Test Suite

All 34 test suites run in under 4.0 seconds with 100% pass rate:

pytest tests -v

============================= 34 passed in 3.97s ==============================

🔒 100% Local & Privacy-First (Zero Telemetry)

We believe developer tools should never spy on you or your AI agents:

• Zero Telemetry: No tracking pixels, no user analytics, and zero remote logging.
• Air-Gapped Safety: All firewalls, anti-thrashing heuristics, AST parsing, and secret redactors run 100% locally on your machine.
• Data Sovereignty: The blackbox flight recorder, antibody memory DB, and shadow sandboxes live strictly in ~/.git_oops/ in your local home directory.
• Credential Masking: High-entropy strings, tokens, and keys are automatically sanitized before reaching any AI model.

🤝 Documentation & Community

• Check out the Help & Beginner's Manual for full MCP step-by-step screenshots and command references.
• Review our Legal Terms & Disclaimer for licensing, privacy, and autonomous agent safety notices.
• Read our Contributing Guidelines to set up your environment.
• Review our Code of Conduct.
• To report security vulnerabilities privately, please review SECURITY.md.