No execution traces detected in the active SQLite database. Execute an instrumented agent pipeline or run the CLI evaluator to stream telemetry.
from tracebackai import trace, TraceContext @trace(step_type="retrieval") def retrieve_context(query: str): return ["passage 1", "passage 2"] with TraceContext(pipeline_name="customer_rag"): docs = retrieve_context("How do I configure telemetry?")
python benchmarks/blame_accuracy.py
None
Select a baseline execution trace and a comparison trace above to inspect step-by-step health score deltas, regressions, additions, or structural divergence.