Add forensic reporting example with agent-forensics

[ilflow4592]

Description

Add an example demonstrating forensic reporting for OpenAI Agents using agent-forensics.

The example shows how to:

• Add forensic hooks to an agent with one line (hooks=f.openai_agents())
• Auto-capture all decisions, tool calls, and LLM interactions
• Generate a forensic report after agent execution
• Auto-classify failure patterns (hallucinated tool output, missing approval, etc.)
• Extract model config for deterministic replay

This is useful for:

• Debugging agent failures ("why did the agent do that?")
• EU AI Act compliance (Article 14 — decision traceability)
• Comparing agent behavior across runs (replay diff)

Usage

pip install openai-agents agent-forensics
python examples/basic/agent_forensics_example.py

Dependencies

• agent-forensics (optional, only needed for this example)

[seratch]

Thanks for sharing this. We don't have this example in this repo but please include it in your own repo!