causal-agent-replay
What it is
A Python library for counterfactual replay and causal attribution over LLM agent trajectories. It models agent runs as structural causal models, allowing users to intervene on specific steps (e.g., resampling actions or forcing specific tool results) and re-run the agent forward to identify which step actually caused a failure.
The problem it solves
Existing observability tools show what happened, and eval tools score pass/fail, but neither identifies which specific step in an agent's trajectory actually caused a bad outcome.
What you could use it for
- Debugging LLM agents that issue incorrect refunds or leak data by identifying the specific decision step that caused the error
- Evaluating the impact of model upgrades or prompt changes on agent outcomes via policy and context interventions
- Generating interactive attribution reports with confidence intervals for agent failures in LangGraph, OpenAI Agents SDK, or CrewAI applications
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.