progress-tracker

You can track continuous improvement loops by querying GitHub issues, pull requests, and CI run health to compute throughput, closure rates, and queue depth over recent periods. The Skill aggregates these signals and appends them to an append-only markdown log.

Pattern detection works by analyzing GitHub issues, pull requests, and CI run conclusions to identify recurring issue types, agent failure trends, CI instability signals, and stale work. It flags these recurring bottlenecks for manual review and meta-improvement.

Yes, you can analyze agent spend logs and CI health together. The Skill reads local agent spend logs alongside GitHub metrics to compute spend efficiency, identify CI instability signals, and detect operational bottlenecks across the last 7 days.

Yes, you need GitHub labels to manage agent-driven issue resolution queues. The Skill applies to teams managing labeled work queues and aggregates closure and success metrics by querying these labeled items across repeated workflows.

The best way to automate meta-improvement issue creation is to run the tracker to detect consistent multi-day signals. When issues repeat across 3 or more days, the Skill conditionally creates limited meta-improvement issues and appends daily log entries.

Your GitHub workflow metrics may not capture recurring bottlenecks without dedicated pattern detection across failures. This Skill resolves the blindness by identifying recurring issue types, agent failure trends, and CI instability signals over repeated 7-day periods.