agent-skills

Produce low-fidelity black-and-white UI wireframes as standalone SVG files. Use when asked to wireframe a screen, page, or flow, make a low-fi mockup or layout draft, bundle wireframes into a browsable viewer page, or publish wireframes for review.

Gate merges and commits of AI-implemented work behind a comprehension quiz that the USER must pass. Use this whenever a nontrivial implementation is complete and the user is about to commit, merge, open a PR, or ship — or whenever the user says "quiz me", "gate this", "do I understand this", or asks to review AI-generated changes before accepting them. Also use proactively at the end of any substantial implementation session, even if the user doesn't ask, offer the quiz before suggesting a commit.

Turn a working session — the docs, issues, and half-formed ideas you've fed in, plus anything you point it at — into a spec and then a set of self-contained, executable implementation plans for OTHER models/agents to build. A senior architect that plans but never implements. Use when you have that context in hand and want it distilled into a PRD-style spec and broken into dependency-ordered, verifiable work items. Publishes to the repo's configured issue tracker or to markdown files.

Audit and harden Node.js projects against npm supply chain attacks — compromised maintainer accounts, malicious package versions, and install-script payloads. Use when reviewing or setting up package.json, lockfiles, .npmrc, Dockerfile, or CI workflows for security; when the user mentions npm security, supply chain attacks, `npm audit`, lockfile policy, install scripts, or min-release-age; also when the user wants to check whether their dependencies are safe, or recover from a suspected compromise.

Manage Val Town projects using the vt CLI. Use when working with Vals (Val Town serverless functions), syncing code to Val Town, creating HTTP endpoints, streaming logs, or managing Val Town branches. Triggers on tasks involving Val Town development, val creation/editing, or when user mentions "vt", "val town", or "vals".

Use when the user gives a rough idea, broad objective, unfinished `/goal`, plan, issue, visual target, optimization target, or research task and wants it refined into a concise Codex Goal contract with one durable objective, clear boundaries, realistic environment, measurable progress, evidence-based completion, checkpoints, cleanup, blockers, and an output strictly under the hard cap of 4,000 characters.

Guides structured ideation through Socratic questioning to explore problems, opportunities, and solutions. Use when brainstorming features, exploring use cases, or thinking through new ideas.

Reviews code changes for reuse, composition, codebase consistency, and slop. Use when asked to review PRs/diffs, check code reuse, composition, cleanliness, or whether code fits the codebase.

Build single-file HTML tools — self-contained HTML+JS+CSS applications that solve a specific problem without a build step. Use this skill whenever the user asks to build a utility, converter, viewer, debugger, or any small interactive web tool. Also trigger when the user says "build me a tool that...", "make a quick app for...", "I need a single-file HTML page that...", or wants to create something that could be hosted as a static file. This skill is specifically for practical, utility-focused tools (not landing pages, portfolios, or marketing sites — use frontend-design for those).

Use before implementing logs in a medium to large scale production system.

Generate comprehensive implementation guides for coding tasks instead of writing code directly. Use when the user requests detailed implementation documentation, step-by-step development guides, or when they want to implement features themselves using tools like Cursor. Creates exhaustive guides with background context, architecture decisions, milestones with verification points, and rationale for a "build-it-yourself" workflow.

Guidelines for proper React useEffect usage and avoiding unnecessary Effects. Use when writing, reviewing, or refactoring React components that use useEffect, useState, or handle side effects. Triggers on tasks involving React Effects, derived state, event handlers, data fetching, or component synchronization.