⏳ This skill is pending AI review.
Scores will appear once the review pipeline completes.
delight
Refine how product interactions feel during UI creation or a polish pass. Build or review feedback, celebrations, character reactions, and progress mechanics when relevant; improve meaningful responses without adding gratuitous effects.
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
Not yet listed on ClawHub or SkillsMP
// README
skills
Seven skill libraries for AI coding agents: code, prose, orchestration, interface design, product delight, design research, and QA.
Each library turns judgement into a concrete decision or check: which code to delete, what makes a sentence weak, whether an animation blocks the task, or what a design reference actually proves. They work in Claude Code, Cursor, Codex, and other agents that read skill files.
Install
npx skills@latest add blakecyze/skills
Or, in Claude Code, install any library as a plugin:
/plugin marketplace add blakecyze/skills
/plugin install kanso@cyze
/plugin install mimesis@cyze
/plugin install swarm@cyze
/plugin install flow@cyze
/plugin install delight@cyze
/plugin install yum@cyze
/plugin install litmus@cyze
Each library also stands alone. Its own README covers tool-specific installers and configuration.
Why use it?
- Scoped passes. Name the surface and the outcome; specialist detail loads when needed.
- A check per rule. Every rule can be evaluated against the source. If it can't, it doesn't ship.
- Context on demand. Standing principles stay short where a library needs them. Flow, Delight, and Yum load only for the relevant task, with supporting detail read as needed.
- Computed numbers. Contrast ratios and scale membership come from scripts. A guessed number is a fabricated one.
Reference
kanso
Code. Keeps AI-written output short and honest: delete before you add, earn every line.
| Skill | Does |
|---|---|
| kanso-principles | Standing anti-dilution rules for all code work. Auto-loads. |
| kanso-audit | Code review and pattern analysis of a diff, branch, or codebase |
| kanso-council | The audit run by every agent CLI on the machine, merged by consensus |
| kanso-refactor | Clean up, tighten, and de-bloat code |
| kanso-nuclear | Deep structural audit of a whole codebase |
| kanso-task | Sharpen a rough request before running it, carefully |
| kanso-commit | Stage work and split it into logical commits |
| kanso-pr | Pull request descriptions and branch summaries |
| kanso-context | Create and prune AGENTS.md / CLAUDE.md files |
| kanso-handoff | Packs the session into one file so /clear costs nothing |
| kanso-prompting | Rules for writing better prompts for frontier models |
mimesis
Prose. Strips AI tells from writing meant for people, and compiles token-efficient instructions for machines. This README went through it.
| Skill | Does |
|---|---|
| mimesis-principles | Standing anti-tell rules for all generated prose. Auto-loads. |
| mimesis-human | Audit, rewrite, or generate text that reads as human |
| mimesis-concise | Shorter and denser without losing meaning |
| mimesis-tone | Five named voices: persuasive, practitioner, warm, blunt, plain |
| mimesis-design | Spot and strip machine-made design tells from an interface |
| mimesis-compile | Dense, structured instructions for machine readers |
swarm
Orchestration. Fans work out to parallel agents, picks winners, and delegates execution to cheaper models, all under a hard cost ceiling.
| Skill | Does |
|---|---|
| swarm-principles | Standing rules for when to fan out and when not to. Auto-loads. |
| swarm-explore | Read-only parallel fan-out over a large codebase |
| swarm-plan | Break a large task into parallel workstreams |
| swarm-improve | Audit a codebase for latent work and capture it as a backlog |
| swarm-execute | Run planned work through cheaper models |
| swarm-bulldoze | Best-of-N: try several approaches, keep the winner |
| swarm-refine | Parallel critics attack one artefact from several angles |
Design refinement passes
Each pass accepts optional scope, review or apply, and focus. Use ordinary language or name the pass. Relevant guidance can activate during design work; a broad pass need not change every domain.
| Pass | Purpose |
|---|---|
| flow | Refine layout, type, colour, controls, accessibility, copy, and motion |
| delight | Refine product feedback, celebrations, character reactions, and progress |
| yum | Find, compare, and adapt relevant design references |
Focused skills: flow-motion, flow-access, delight-character, and delight-habits. Use them for dedicated work; routine refinement stays in the broad pass.
Each includes its own resources. Installation and migration notes: Flow, Delight, Yum.
litmus
Behavioural QA with evidence tied to the tested build and environment.
| Skill | Purpose |
|---|---|
| litmus | Plan, run, or repair a scoped web or mobile QA pass |
| litmus-flakes | Diagnose intermittent failures without retrying until green |
Layout
skills/
kanso/ code
mimesis/ prose
swarm/ agents
flow/ design
delight/ product feedback
yum/ design research
litmus/ QA
Each library is a complete Claude Code plugin with its own README, changelog, and licence. The standalone repos (kanso, mimesis, swarm) remain live. This repo is the front door.
Licence
MIT.
// HOW IT'S BUILT
KEY FILES