Intentional UX
Task paths from intention to verified outcome — costs, recovery, evidence levels.
npx skills add actually-useful-ai/intentional-ux
Portable Agent Skills CLI — installs into Cursor, Claude Code, Codex, and other supported agents. Source: actually-useful-ai/intentional-ux.
Native marketplace CLIs (optional): claude plugin marketplace add actually-useful-ai/intentional-ux; Cursor: cursor-agent plugin marketplace add https://github.com/actually-useful-ai/intentional-ux.
Install
npx skills add actually-useful-ai/intentional-ux
claude plugin install intentional-ux@intentional-ux
cursor-agent plugin marketplace add https://github.com/actually-useful-ai/intentional-ux
Claude Code slash commands. Cursor and Codex paths are on the GitHub README when available.
person/input mode -> start -> outcome -> success evidence
Features
Task tuple
Name the person, start, outcome, and observable success evidence.
Metric vector
Completion, hits, Fitts, cognitive load, wait, recovery, surface, a11y, integrity — no fake combined score.
Evidence levels
Measured, Observed, Assumed — unknowns stay visible.
Capability paths
Agent skills have task paths too: discovery, resolution, activation, fallback.