otel-validation

## Context Use when auditing Squad usage, delegation quality, model mix, tool-call behavior, or session bottlenecks. ## Patterns - Prefer telemetry/session-store evidence when available. - Significant work-unit routing accuracy is the primary KPI; raw tool/task counts are supporting evidence. - Pair telemetry with `.squad/orchestration-log/` entries and git history. ## Examples - A drift audit checks whether substantial code changes had corresponding agent delegation and orchestration logs. - A model-policy audit checks whether routine work stayed on standard/fast models instead of premium reviewers. ## Anti-Patterns - Do not judge delegation quality only by raw tool counts. - Do not expose private telemetry contents in user-facing summaries.