Generate a health assessment of the design system — source parity, token coverage, contribution velocity, adoption signals, and delivery outcomes.
Read context/ds-context-guide.md for environment detection and configuration. Read context/output-patterns.md for output conventions.
Generate a health assessment synthesizing available data: token coverage, component adoption signals, source parity (are design and code in agreement?), contribution velocity, open issues, and accessibility status. Outputs a leadership-ready narrative focused on delivery outcomes, not just adoption numbers. The best systems are quietly embedded in "how we work here" — this skill surfaces that invisible value.
$ARGUMENTS = optional focus area (e.g., "tokens", "adoption", "accessibility") or time period. Default: full health check, last 30 days.
Detect available sources per ds-context-guide.md. This skill benefits from many integrations but works with any subset.
Gather data from available sources. Launch subagents in parallel where possible:
For any source not available, note it in the output and work with what you have. If no MCPs are available, produce a qualitative assessment from codebase analysis alone; note which dimensions lack data.
Synthesize into health dimensions:
Translate to delivery outcomes — not "47 components adopted" but "47 components in active use across 12 product teams, estimated to save X hours of duplicated implementation effort."
Follow the report template from output-patterns.md. Include: