What problem does it solve? Operators of an autonomous agent fleet must open multiple report files (skill-health analytics, heartbeat logs, tweet-allocator totals, repo-pulse deltas) to assemble a weekly picture of whether the agent's work was worth it. This Skill consolidates those scattered signals into a single scorecard, ops day-recap, or push deep-dive with a clear OK/WATCH/DEGRADED verdict. ## Core Features & Use Cases - Weekly operator scorecard: Synthesizes agent health, community growth, and economic activity from existing skill output files into a three-paragraph verdict article, dashboard JSON spec, and notification. - Ops day recap: Summarizes one day of agent activity — what shipped, what failed, what needs a human decision — with deduplication, silent-failure detection via cron-state, and a mandatory TL;DR. - Push recap: Reads commit diffs and merged PRs from watched GitHub repos, classifies commits as user-visible/internal/infra, ranks impact, and suppresses quiet days. - Use Case: On Monday morning, run the default scorecard mode to receive a notification with the week's fleet success rate, star/fork growth, $AEON distribution totals, and a worst-of-three verdict, without opening four separate report files. ## Quick Start Run the operator-scorecard skill with an empty argument to generate this week's scorecard, or pass 'ops' for a daily operational recap or 'push' for a commit-level push recap.
