Some checks failed
Forge CI / smoke-and-build (pull_request) Failing after 1m7s
Cron job run_job() ran the agent but never extracted token counts from the result dict. Usage data was accumulated by the agent during the conversation, then discarded on return. Fix: Add _format_token_usage() helper that extracts input_tokens, output_tokens, total_tokens, api_calls, and estimated_cost_usd from the agent result. Include a Token Usage section in all three output templates (success, script failure, sync fallback). Closes #490