AY Skills

AY Skills is a set of fifteen standalone skills for coding agents. Each skill defines when the agent can proceed, when a decision needs your approval, and what evidence counts as done.

Quick start

List the skills, then choose what to install:

npx skills@latest add ALVIN-YANG/ay-skills -l
npx skills@latest add ALVIN-YANG/ay-skills

Describe the task normally. The matching skill should load from the request; naming $ay-implement or another skill is optional.

Add CSV export to this app and verify the result.

How it behaves

A precise request runs directly.
A choice that changes product behavior, architecture, data contracts, dependencies, scope, risk, cost, rollback, or an external action stops for approval.
Approved work continues inside that boundary and reports the highest proof actually reached.
Reviews stay read-only unless fixes are requested.

The skills are independent. They do not require a router, shared runtime, modes, hooks, telemetry, automatic commits, or planning files.

Skills

Plan and design

Skill	Use it for
`ay-expert-lens`	Apply a real expert's documented framework to an open decision
`ay-product`	Research a product idea and define a focused, testable direction
`ay-ui`	Choose a visual direction, design screens, and prepare a UI handoff
`ay-architecture`	Set system boundaries, ownership, deployment, and failure behavior
`ay-api`	Design consumer-facing APIs, events, webhooks, and service contracts
`ay-database`	Design schemas, indexes, transactions, and migrations for the exact engine

Build and hand off

Skill	Use it for
`ay-implement`	Implement an approved change and verify the result
`ay-integration-docs`	Document only the API and message changes a specific client needs
`ay-fix`	Confirm a bug's cause, then make the smallest useful repair
`ay-improve`	Refactor or optimize from a measured baseline
`ay-review`	Review product and delivery artifacts without silently editing them

Specialized work

Skill	Use it for
`ay-audio`	Build or debug streaming TTS and voice playback on Apple platforms
`ay-write`	Write or substantially revise articles, tutorials, and visual explainers
`ay-icon`	Design and install store or launcher icons
`ay-store-screenshots`	Produce reproducible App Store screenshots from shipped UI

See Choosing a skill for common overlaps and a typical product handoff.

One example

You: Add offline mode.

AY: I found that the app currently assumes a live connection. I recommend making
cached data read-only offline, showing when it became stale, and leaving queued
writes out of this change. I will check cold launch, reconnect, and stale-data
behavior. Approve this boundary?

You: Approved.

AY: Implemented the approved flow and ran the focused tests. Queued writes remain
out of scope; real-device reconnect behavior is the only unverified surface.

Evidence in this repository

The test suite checks routing, behavior, protected files, reproducible assets, portable installation, release archives, and multi-skill handoffs. The three end-to-end journeys—FieldLog, PairDown, and NextStop—are synthetic fixtures, not claims of user research or production adoption. Their inputs and assertions live in tests/journey-scenarios.json.

Development commands and the limits of each proof layer are in CONTRIBUTING.md.

Other install options

Install all fifteen skills for Codex and Claude Code:

npx skills@latest add ALVIN-YANG/ay-skills -s '*' -a codex claude-code -g -y

/plugin marketplace add ALVIN-YANG/ay-skills
/plugin install ay-skills@ay-skills

npx skills@latest list -g
npx skills@latest update -g
npx skills@latest remove ay-audio -g

AY Skills follows the Agent Skills specification. The Codex manifest is for local marketplace testing and a possible future directory submission; this repository does not claim a public Codex plugin listing.

Acknowledgements and license

The project is original work informed by several agent-skill and app-delivery projects. The exact sources, licenses, and boundaries are recorded in docs/influences.md.

Ay Skills

At a glance

README