CHOOSE BY OUTCOME
What do you need to understand? Seven complete, inspectable workflows 07 · INFORMATION DESIGN Can readers find and understand the answer? Apply F, layer-cake, spotted, and comparison-table patterns while preserving a path to close reading. Explore Scan to Comprehend ↗ 01 · WEB CLARITY Can people and agents use this page? Audit its answer, evidence, semantics, structured data, decision support, and honest limits. Explore Answer-Ready Web ↗ 02 · RELEASE CONFIDENCE Is this release ready? Review change scope, contributors, tests, security, supply chain, and operational evidence. Explore Release Brief ↗ 03 · DATA QUALITY Can I trust this CSV? Find missing values, type conflicts, duplicates, category drift, and structural risk. Explore CSV Profile ↗ 04 · AGENT COORDINATION Who needs to know? Route a change to relevant active sessions, then distinguish selection from delivery. Explore Message Routing ↗ 05 · INTERFACE JUDGMENT Why does this interface feel difficult? Connect observable evidence to a small set of UX principles, then make every recommendation testable. Explore Laws of UX ↗ 06 · DESIGN SYSTEM How should this interface look and behave? Generate semantic tokens, component states, layout, responsive rules, guardrails, and an agent-ready contract. Explore Interface Design System ↗
BROWSE THE LIBRARY
Ready workflows Original Triuna Labs collection · October 2026 ⌕ AgentAll agents Codex Claude Code Other agents CapabilityAll capabilities
Loading skills…
Reset filters
No matching skills Try another search or reset the filters.
We couldn’t load the catalog. Browse skills on GitHub or reload this page.
Enable JavaScript for search and filters. Browse the skill instructions below:
Make a page answer-ready — Audit or create a landing page that clearly answers buyer questions, exposes crawlable semantic content, and ties consequential claims to visible evidence. Use for product, service, consulting, or open-source landing pages and for SEO/AI-search readiness reviews; does not promise rankings or citations.Create an interface design system — Generate, validate, or evolve a coherent interface design system from a product brief, existing brand evidence, or interface constraints. Use for semantic color and type tokens, component states, layout, elevation, responsive behavior, guardrails, agent prompts, and an HTML specimen; not for logo creation or unsupported claims about user outcomes.Profile a CSV — Analyze a local CSV for structural and value-quality risks, then generate interactive HTML, Markdown, or JSON reports with optional aggregate-only Laya semantic review and confidence gates. Use before importing, modeling, or cleaning a CSV; does not repair records automatically.Review an interface with UX principles — Review, compare, or constrain an interface using evidence-backed UX principles, and render an annotated teardown when useful. Use for decision load, target acquisition, grouping, expectations, complexity, feedback, progress, and endings; not for visual tokens, CSS architecture, or standalone WCAG conformance.Build a release brief — Build an evidence-backed release brief from an explicit Git revision range, with collaboration context, quality and security deltas, supply-chain and operational readiness, and audience-aware Markdown, JSON, interactive HTML, or print-to-PDF reports. Use when preparing release communication or reviewing release readiness; does not tag, deploy, publish, or infer that merged work shipped.Route an agent message — Route a change or question from one working session to the relevant active agents or sessions, with per-recipient relevance decisions and accurate delivery receipts. Use when coordination spans concurrent work; does not manage project memory or assume that a selected recipient received or acted on a message.Scan to Comprehend — Design or audit information-dense websites, articles, and infographics using task-aware scanning patterns. Includes source, assistive, crawler, retrieval, and experimental attention views; does not measure gaze or comprehension.“Format-compatible” means a portable package, not a verified agent run. Other-agent runtimes are untested. Review requirements before installing.
Standalone skill: Supply Chain Graph ↗ maps selected project folders and dependency relationships. Its local dashboard and scan output stay outside this catalog.
