royvergara
UserThe open source operating system for design teams running on AI — three gates, gate-enforcing Claude skills, a work ledger, and a conductor. Installable as a Claude Code plugin.
Categories
Indexed Skills (18)
brief-from-pain
Use when turning a validated customer pain and a first-pass IA into a design brief, before any prototype prompt. Triggers on a named pain plus a request for a brief or "what should we build." Will not write a brief until success is defined in advance; an unvalidated pain proceeds only as an explicit owned bet, and the bar is never excused.
brief-to-prompt
Use when converting a design brief into a prompt for an AI builder — v0, Bolt.new, Lovable, Replit, Claude Artifacts, Framer, or any generator. Triggers on a brief plus a request to write a prototype or build prompt for a named tool. Will not output a prompt until the brief defines what good looks like; full-app builders also need the data settled first.
conductor
Use when someone needs to know where a piece of work stands and what can run next — resuming after a gap, handing work to someone else, entering mid-stream with artifacts already in hand, or asking "where are we" / "what's next" on design work. Reads state from a design-os.work ledger or from whatever is described in hand. Routes only — it never judges a gate itself, and it refuses to count a bare checkmark or a verbal "that's done" as a passed gate, because only the artifact carries state.
critique-synthesis
Use when synthesizing scattered critique from multiple reviewers into a direction. Triggers on pasted feedback from two or more people plus a request to make sense of it. Returns a ranked decision with the strongest signal named, not a summary.
design-system-enforcement
Use when auditing generated or drafted UI against a team's design system. Triggers on a screenshot, code, or spec plus a reference to the design system. Returns violations, not compliments.
outcome-readout
Use after a shipped feature has run long enough to read its analytics, to judge whether it solved the pain and name the next thing worth building. Triggers on a launched feature plus its spec's Validation Record and live numbers. No pre-registered metric and measured value, no verdict.
outcomes-scorecard
Use to render a filled AI Outcomes Scorecard into a self-contained, shareable HTML page. Triggers on a filled ai-outcomes-scorecard.md plus a request to render, publish, or share the scorecard as a page or artifact. Renders earned state only — it never judges a feature and never routes; no baseline, no render, and a leverage number never renders as a result.
period-review
Use to close out a period — a quarter by default — and render the program's judgment record across every closed ledger in design-os.work/. Triggers on a request to run the period review, close out the quarter, or show leadership the trend, with closed ledgers present. A first period with no prior frozen review renders as the baseline, never a trend; judgment below n=10 calls renders as counts, never a percentage; and coverage — how many efforts shipped against how many ran through a ledger — always prints, even when the gap is unflattering.
prd-to-ia
Use when turning a PRD or requirements document into a first pass information architecture. Triggers on a pasted or linked PRD plus a request for an IA, sitemap, screen list, or structure. Always returns what was excluded and why, not just the IA.
prototype-to-spec
Use when turning a chosen prototype into a buildable spec for engineering handoff. Triggers on a prototype link, file, or description plus a request for a spec or handoff. No validation signal, no spec — the only exception is an explicit owned bet (named owner, declared absence of evidence, reason, review date), which produces a spec with a Bet Record instead of a Validation Record.
prototype-triage
Use when a generated prototype is a candidate for the team's review time, to triage it against its brief first. Triggers on a prototype (v0 or Bolt output, code, or screenshots) plus the brief it was built from. Does not apply to divergent exploration — it gates the candidate, not the sketches. No brief, no triage; and never a thumbs-up.
research-to-pain
Use when turning raw research (interview notes, support tickets, sales call logs, analytics, survey responses) into a small set of validated customer pains, before any IA, journey map, or brief. Triggers on a pile of research plus a request to find the real problem or "what's the pain here." Will not promote a pain to validated on stakeholder opinion, a feature request, or a single untriangulated source.
team-ai-baseline
Use when you need an honest read of where a design team actually sits against its AI mandate, before you plan any AI initiative around it. Triggers on a request to baseline a team, assess AI adoption, locate a team on the maturity curve, or diagnose why a mandate is not landing. Refuses to count tools bought or intentions stated as adoption, and will not place a team a stage above what its real working practice can support.
validation-plan
Use when you need the smallest test that would settle a decision — validate a weak pain, earn a signal for a prototype direction before a spec, or design the read a shipped feature will be scored against. Triggers on a thing to validate plus "what's the smallest test" / "how do we validate this" / "design the test." Will not design a test with no decision behind it.
weekly-review
Use to prep the weekly gate review — the agenda of what moved, what stalled, what needs a decision, and what can run now, read from the open work ledgers. Triggers on "run the weekly review," "prep the gate review," or "what moved this week," with ledgers or described state in hand. It surfaces and routes; it never judges — no gate gets certified, softened, or flipped in a meeting agenda, and with no state to read it refuses to fabricate a review.
design-system-extraction
Use when producing a design-system reference from a team's existing components, tokens, and patterns — the reference design-system-enforcement needs to audit against. Triggers on existing UI source (a component file, a codebase, exported tokens) plus a request to extract, document, or produce the design system. Returns an evidence-backed reference with its conflicts named, never an invented system.
figma-plugin-orchestration
Use when a design task spans multiple Figma plugins or steps and needs to be sequenced from one instruction. Triggers on a request to coordinate, sequence, or automate plugin work in Figma. Always names which steps stay human.
user-journey-mapping
Use when drafting a user journey map from a brief or research inputs. Triggers on requests for a journey map, experience map, or end to end flow of a user's experience. Will not generate a map without a named customer pain and at least one real signal behind it.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.