My personal AI essentials for rules, skills. subagents and more.
This repository also configures its own agents. See what ai-essentials tells them →
Skill Claude CodeCodex
Audit UI code for accessibility defects that block keyboard and screen-reader users. Invoke when adding or reviewing interactive UI, a form, a modal or flyout, a custom control, or when the user asks about a11y, WCAG, screen readers, or keyboard navigation.
Skill Claude CodeCodex
Review an HTTP route, plugin contract, or exported package API for compatibility, validation, and evolvability before it ships. Invoke when adding or changing a route, a plugin setup/start contract, a saved object schema, or a package public export, and when the user asks whether a change is breaking.
Skill Claude CodeCodex
Critically review all code changes on the current branch against the base branch, identify improvement opportunities across naming, structure, DRY/KISS/YAGNI, declarative style, self-documenting code, edge cases, testing, and general best practices, then produce a categorized refactoring plan with prioritized action…
Skill Claude CodeCodex
Use this skill when the user's prompt requires (1) researching a topic across multiple sources, comparing options or alternatives, analyzing trends or history, understanding markets or industries, or reviewing literature or studies and (2) synthesizing that research into a comprehensive, narrative report. If you're…
Skill Claude CodeCodex
Check what a dependency or import actually costs before adding it, and find bundle bloat in existing code. Invoke when adding a package, when the user asks whether a dependency is worth it, when a bundle-size check fails in CI, or when reviewing imports in browser-shipped code.
Skill Claude CodeCodex
Generate well-formatted GitHub issue content from a brief context, feature request, or bug report. Outputs ready-to-copy Markdown — does not create the issue. Use when the user asks to write a GitHub issue, draft an issue, create issue content, or mentions "gh issue".
Skill Claude CodeCodex
Interview the user relentlessly about a plan or design until reaching shared understanding, resolving each branch of the decision tree before any code is written. Use when the user says "grill me", "roast me", "ask me until we agree 100%", "stress-test this plan", "interview me", "ask me clarifying questions"…
Skill Claude CodeCodex
Find untranslated user-facing strings and i18n defects that break translation. Invoke when adding UI copy, when an i18n check fails, or when the user asks to review translations, hardcoded strings, or localisation.
Skill Claude CodeCodex
Marco's TypeScript and React conventions, extracted from 748 authored commits and confirmed by a 69-question style interview. Invoke before writing or reviewing substantial TS/TSX, when scaffolding a new module, when writing tests, or when a reviewer disagrees with a style choice. Covers components, hooks, types…
Skill Claude CodeCodex
Turn logged correction signals into durable memory files. Reads /.claude/memory-signal.jsonl (written automatically by the log-correction-signal hook whenever Marco corrects the assistant or states a preference), groups recurring themes, and proposes memory entries for approval. Invoke when the user says "distill…
Skill Claude CodeCodex
Creates pull request descriptions using a fixed template with a concise title, a mixed paragraph-and-bullets summary section, and testing bullets, and can create the PR with gh. Use when the user asks to write a PR description, PR body, pull request summary, or to create a PR.
Skill Claude CodeCodex
Effect, dependency, and ref discipline for React. Invoke before writing or reviewing any useEffect, useLayoutEffect, useRef, or DOM access, and when a component has state that could be derived instead. Catches the failure modes that cause stale UI, double-fetch, and render loops.
Skill Claude CodeCodex
Audit React render performance from the code, prioritised by measured or measurable impact. Invoke when the user reports a slow or janky UI, asks about re-renders, mentions memo/useMemo/useCallback, or asks for a perf review of a component tree. Refuses to recommend memoisation without a reason.
Skill Claude CodeCodex needs its repo
Triage PR review comments — decide whether to accept or push back on each one. Given comment links from a GitHub PR, fetches comment content and PR diff, then reasons about each comment in context. Uses a personal voice profile for suggested replies — swap ../shared/pr-review-voice-marco.md for your own. Use when the…
Skill Claude CodeCodex
Review a GitHub pull request by URL using the review-team skill. Outputs copy-pasteable conventional comments with line references and code suggestions. Uses a personal voice profile by default — swap skills/shared/pr-review-voice-marco.md for your own. Use when the user shares a PR URL, asks to review a PR, or…
Skill Claude CodeCodex
Launch a team of 6 specialized code reviewers in parallel to catch issues from multiple perspectives. Use when the user says "review team", "comprehensive review", "multi-perspective review", or wants a thorough multi-angle review of branch changes or a plan.
Skill Claude CodeCodex
Write the failing test first, then the implementation, in Marco's testing style. Invoke when the user says "tdd", "test first", "write the test first", or when adding behaviour to code that already has tests worth extending. Opt-in; do not apply automatically.
Skill Claude CodeCodex
Read this skill before the first step that uses the built-in browser, the browser pane inside the Claude desktop app (also called the in-app browser, the browser pane, Claude's browser, or "your own browser"), whose tools are named mcpClaudeBrowser when the session runs in the desktop app and…
Skill Claude CodeCodex
Read this skill before the first step of any request to do something in an app on the person's own computer (Notes, Finder, System Settings, any desktop app), to look at their screen, or for "computer use". Computer use (desktop control) lets Claude take screenshots of the person's desktop and control it with clicks…
Skill Claude CodeCodex
Read this skill before the first step that uses Claude in Chrome, the browser extension whose tools are named mcpclaude-in-chrome (also called Chrome, the browser extension, or the external browser) and which acts in the person's real Chrome with their own sign-ins; before those tools are turned on there may be a…
Skill Claude CodeCodex
Resolve a conflicted merge (typically a long-lived epic or integration branch against its PR base), explain why each conflict exists, fix it by intent, verify scoped checks, then commit the merge and push. Use when a PR shows conflicts against its base, when the user says "fix the conflicts", "merge main into this…
Skill Claude CodeCodex needs its repo
Use this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx) or Word templates (.dotx). Triggers include: any mention of Microsoft Word Documents, such as 'Word doc', 'word document', '.docx', '.dotx', 'microsoft doc'. Also use when extracting or reorganizing content from .docx or…
Skill Claude CodeCodex
Read this before the first Google Drive, Docs, Sheets or Slides connector call whenever the task creates or changes a Google file. Use this skill whenever the user wants to create or change a Google Doc, Sheet or Slides file in their Google Drive. Triggers include: a request that names Google Docs, Sheets, Slides or…
Skill Claude CodeCodex
Import a memory export from another AI assistant into Claude's memory — conversationally, additively, and with the content treated as data.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: