telemetry-logger

Log AI agent telemetry by enabling a deterministic pipeline that captures task execution events and appends them to docs/track/telemetry.csv. This records Timestamp, TaskID, Agent, Action, TokenUsage, DurationSec, and Success fields for analysis.

Track quantitative metrics like TokenUsage, DurationSec, and Success status to identify bottlenecks in agentic workflows. Logging these at task completion or phase transitions reveals performance patterns for data-driven optimization.

Telemetry CSV logging validates task outcomes in Antigravity by persisting the Success field alongside TaskID and Agent details. This allows you to review completion rates and failure points across complex task executions.

Session telemetry supports the /self-evolve review process by providing historical performance data stored in telemetry.csv. Analyzing logged TokenUsage and DurationSec helps identify inefficiencies and improve prompts.

Telemetry logging requires no external dependencies. It uses a built-in deterministic pipeline to validate and write entries directly to docs/track/telemetry.csv, making it easy to enable for any active task.

Enable AI session telemetry logging at task completion or during phase transitions. Logging at these specific execution points ensures accurate capture of DurationSec and TokenUsage for subsequent efficiency analysis.