If you're building GitHub workflow agents or dashboards, this gives you the scoring math to turn messy repository data into actionable numbers. It includes a 100-point health score with specific penalty weights for failing CI, stale PRs, and missing community docs, plus priority scoring formulas for issues and PRs based on mentions, labels, and activity. The confidence levels (High/Medium/Low) and delta tracking (Fixed/New/Persistent/Regressed) are smart additions that keep your agent from crying wolf. The bottleneck detection categories are concrete enough to actually implement. It's essentially a reference card for turning GitHub API responses into scores and grades instead of reinventing the weights every time.
npx skills add https://github.com/community-access/accessibility-agents --skill github-analytics-scoring