speckit-token-observability

To track token usage across Spec Kit workflow phases, you need a token observability process that captures data for specify, plan, tasks, and implement phases. It normalizes run labels and writes metadata, step records, and summaries into a token-analysis directory.

Yes, you can check token budgets before approving Spec Kit PR automation by applying phase-level reporting. This verifies token costs per phase, checks budget status, and generates required quality notes for final outputs.

If the token-analyzer extension is missing during a workflow run, the system handles the missing data with explicit warnings. It applies guardrails to ensure missing token-analyzer support is reported rather than skipping analysis silently.

You can record token costs for custom orchestration phases like intake, slicing, scheduling, integration, review, commit, and PR creation by organizing each run with a stable label and capturing phase-by-phase token data.

Yes, multi-agent planning requires run-label normalization for accurate token reporting. Normalizing run labels ensures token usage metadata and step records are consistently organized across all Spec Kit and custom orchestration phases.