🚀 Agent Blame-Finder

"Find out which Agent messed up in 3 seconds."

---

🧐 What is this?

When your Multi-Agent system breaks down, it usually goes like this:

• Agent A says: "Agent B gave me bad instructions"
• Agent B says: "I just followed Agent C's output"
• Agent C says: "I just passed data, not my problem"
• You: Digging through thousands of log lines, finding nothing

Agent-Blackbox solves this.

It installs a "Black Box" in every Agent that automatically records:

• Who made the decision
• Which parent output this decision was based on
• When it happened
• Cryptographic signature to prevent denial

When something breaks, just enter the incident hash — 3 seconds later, you know exactly who to blame.

---

✨ Core Features

Feature	Description
One-Click Blame Analysis	Input incident ID, instantly locate the broken link in your responsibility chain
Causality Tree Visualization	Git-like tree showing your Agent decision chain
Cryptographic Signatures	Every decision gets an Ed25519 signature. No denial possible.

---

🚀 30-Second Quick Start

Installation

pip install agent-blame-finder

Usage

from blame_finder import BlameFinder

# Initialize the black box
finder = BlameFinder(storage="./logs")

@finder.trace(agent_name="Coder-Agent")
def write_code(requirement: str) -> str:
    # Your Agent logic here
    return "print('hello world')"

# Execute
result = write_code("write a hello world")

# When something breaks, find out who to blame
print(finder.blame(incident_id="task_123"))

Sample Output

{
  "incident": "task_123",
  "verdict": "Coder-Agent",
  "reason": "Input requirement was correct, but output didn't match expectations",
  "chain": [
    {"agent": "PM-Agent", "action": "Dispatch task", "status": "ok"},
    {"agent": "Coder-Agent", "action": "Write code", "status": "failed"},
    {"agent": "Verifier-Agent", "action": "Verify", "status": "not_reached"}
  ],
  "confidence": 0.94
}

---

🖥️ Visual Dashboard

blame-finder dashboard

Open your browser and see:

• Causality Topology — Who called whom, at a glance
• Failed Nodes Highlighted in Red — Problem agents stand out
• Blame Share Pie Chart — PM Agent takes 60% blame, Coder Agent 35%, Others 5%

---

📦 Supported Languages / Frameworks

Language / Framework	SDK Status
Python	✅ Available
TypeScript	✅ Available
Rust	✅ Core implemented
LangChain	🚧 In development
CrewAI	🚧 In development

---

📅 Roadmap

• Rust core engine
• Python/TypeScript SDK
• LangChain official adapter
• CrewAI official adapter
• One-click PDF/HTML report export

---

🤝 Contributing

PRs welcome! Whether it's fixing bugs, improving docs, or adding features.

1. Fork the repo
2. Create your feature branch (git checkout -b feature/amazing)
3. Commit your changes (git commit -m 'add amazing feature')
4. Push to the branch (git push origin feature/amazing)
5. Open a Pull Request

---

📜 License

MIT © Agent Blame-Finder