agent-skills

Systematic implementation using APEX methodology (Analyze-Plan-Execute-eXamine) with parallel subagents and self-validation. Use when implementing features, fixing bugs, or making code changes that benefit from structured workflow.

Use this skill whenever a user has an audio file (.wav/.mp3/.flac/etc.) that needs to loop as background on a website or web page — hero ambience, landing-page atmosphere, portfolio mood audio, ambient wind/rain/ocean/forest/breeze/pad beds, or any "play quietly behind the page while users read/scroll/interact" use case. Trigger on intents like making a clip loop without audible gaps or boundary artifacts on a site, fixing a click/tick/pop/bump at the loop boundary, fixing stereo bias or L/R imbalance on a web-bound ambient clip, normalizing loudness (LUFS) for web delivery, encoding a loop for web playback, or fading audio in on first user click/gesture — in any web framework (Next.js, Vue, Astro, Nuxt, plain HTML). The skill outputs a gapless FLAC plus a paste-in Web Audio snippet that unlocks on first interaction. Skip for general audio editing (cuts, mixing, effects), music/podcast mastering, or transcription.

Govern the BRAND-VOICE.md — extract a brand's writing voice from URL, Notion page, MD file/directory, or interview into a structured executable doc; update incrementally; diff versions; validate the canonical format; show testable rules. Supports multi-voice via `voice.extends` (founder on corporate, persona on institutional, multi-host) with `_replace`/`_remove` overrides. Consumed by writing skills via `-f` (e.g. `humanize-en`). Use when defining, extracting, refreshing, validating, or inspecting a brand's writing voice — "create a brand voice doc", "extract voice from this site", "tone of voice", "writing style guide", "BRAND-VOICE.md", "founder voice", "persona voice", "multi-voice".

Eight-axis judgment code review for the current diff — Correctness, Simplification, Tests, Documentation, Style, Intent, Design/API, Performance (+ Coherence on metadata changes). Five-phase pipeline scope → deterministic tool battery (npx/uvx-preferred, zero-install for the JS + Python majority) → 8 parallel LLM axis reviewers → Haiku validators on sub-80 findings (verbatim rubric, ≥80 threshold) → synthesis with no-silent-drop + Conventional Comments JSONL. Every report closes with "What I did NOT check" (security → /security-review, runtime perf, flaky detection). Opt-in flags `--verify-build`, `--mutation-test`, `--reconcile`, `--apply-safe`. Public-skill posture — zero auto-install, graceful skip on missing native tools. Invoke before a commit or PR — "review my changes", "deep review", "did I miss anything", "check before I commit", "audit this PR", "drift / gaps / blind spots".

Govern an existing DESIGN.md — Google's open standard for design tokens (YAML frontmatter + eight prose sections). Auto-activates during UI edits to enforce token-only sourcing for colors, typography, spacing, and corner radius when a DESIGN.md is present; steps aside when none exists — it never forces or authors one (`/award-design` authors it up front, then builds under it). Exposes seven CLI-backed subcommands — audit (lint + fix), diff (regression check), export (Tailwind / DTCG), spec (canonical spec), migrate (port legacy Stitch), init (minimal scaffold fallback), audit-extensions (drift check vs globals.css @theme). Token-affecting UI changes update DESIGN.md first, audit, then propagate to code.

Download video or audio from ~1800 sites (YouTube, Vimeo, SoundCloud, Twitch, X…) with the `yt-dlp` CLI — full videos, mp3 audio, playlists, time-range clips, subtitles, capped resolutions. Use whenever the user wants to download, save, grab, or rip a video/audio/playlist from a URL, extract a song as mp3, clip a segment, or fetch subtitles — even when they just say "download this", "get the audio", or "save this playlist".

Pre-implementation thinking — research the problem space, weigh approaches with devil's-advocate rigor, decide what's reversible and conventional while surfacing the load-bearing forks for the user, and emit one apex-ready plan. Use whenever a non-trivial task needs thinking, comparison, or decomposition before code — even when the user doesn't say "forge" (e.g. "should we", "what's the best way", "compare", "evaluate", "pros and cons", "think through", "plan this", "break this down", "spec out", "create issues for", "map out the steps"). Produces a Decision by default (terminal — discuss-then-build); promotes to a Spec with prioritized workstreams when the build path is clear. NOT implementation (→ /ultrapex, /apex, or /oneshot).

Convert any document to Markdown with Microsoft's `markitdown` CLI — PDF, Word, Excel, PowerPoint, HTML, CSV, JSON, XML, ZIP, EPub, images (OCR/EXIF), audio (transcription), and YouTube URLs. Use whenever the user wants to extract text from a binary document, transcribe audio, OCR an image, scrape a YouTube transcript, or pre-process a file for an LLM context window — even when they just say "convert this pdf", "what's in this docx", "transcribe this mp3", or "get the text out of this".

Bootstrap a new web project on a strictly opinionated Cloudflare Workers stack — Next.js 16 or Astro 6, TypeScript strict, pnpm, Biome, Tailwind. No fallbacks: no Vercel/Netlify, no ESLint/Prettier, no swap. Skip if the user wants any of these. Use when the user says "start a new project", "bootstrap", "init", "scaffold", "create a new site", or is working in an empty directory and wants production-ready foundations.

Outcome-driven end-to-end implementation for frontier models — scope the task, commit to an approach, build, and adversarially verify with parallel subagents, in one run. Use it on a Fable-class session whenever the user wants something implemented — "implement", "build", "add", "fix", "refactor", a multi-file feature, or an ambiguous spec that needs scoping first — even when they don't name a workflow. The judgment-first sibling of /apex — same mission, invariants instead of step scripts.

Create a looping background video — crossfade the loop point, encode optimized MP4 + WebM, optional poster frame. Use whenever the user has a video that needs to loop for web use (hero sections, backgrounds, landing-page ambience), wants to fix a visible jump at the loop point, or needs to optimize a clip for web delivery — even if they just say "make this loop smoothly" or "compress this for the site".

Create, configure, and orchestrate Claude Code subagents — specialized Claude instances with focused roles and limited tool access. Covers YAML frontmatter (name, description, tools, model, permissions, hooks, MCP servers), system prompt design, tool restriction patterns, background execution, and multi-agent orchestration. Use whenever the user mentions subagents, delegation, specialists, agent configs, `.claude/agents/`, the `/agents` command, or wants to parallelize work — even when they just say "background agent" or "delegate this".

World-class frontend design engineer for award-winning websites (Awwwards SOTD 7.5+, FWA, CSSDA). Takes the lead on frontend design and build — forces a committed, anti-default visual universe, writes it as a DESIGN.md, then builds the frontend itself under that direction with real assets, premium motion, and anti-AI-slop discipline. Adapts to an existing DESIGN.md and alerts when it is thin. A review mode audits any site against awwwards criteria and anti-slop at any time. Frontend only — routes single-token tweaks to design-system, never touches backend. For landing pages, portfolios, product and marketing sites, and redesigns — not dashboards or internal tools.

Create and optimize CLAUDE.md memory files or .claude/rules/ modular rules for Claude Code projects. Covers file hierarchy, content structure, path-scoped rules, best practices, and anti-patterns. Use when working with CLAUDE.md files, .claude/rules directories, setting up new projects, or improving Claude Code's context awareness — even when the user just says "memory file" or mentions Claude instructions without naming the filename.

Strip AI writing tells from English prose — em-dash overuse, rule of three, negative parallelisms, AI vocabulary (delve, tapestry, crucial, pivotal, underscore, showcase), vague attributions, promotional tone, conjunctive padding (moreover, furthermore, indeed), hedging, signposting, chatbot artifacts. Preserves meaning, structure, code blocks, links, anchors, and frontmatter — rewrites only the flagged phrasing. Operates on inline text or a prose file path; optionally loads a `BRAND-VOICE.md` via `-f` for brand-specific rules. Use whenever English prose needs to sound less machine-generated — "humanize this", "remove AI tells", "clean up the AI slop", "sounds like ChatGPT", "polish the English" — on READMEs, docs, release notes, blog drafts, PR bodies, or commit messages.

Notion access from Claude Code via the official MCP connector (default path, ~95% of intents) or the `ntn` CLI (only for file uploads, Notion Workers, headless/CI scripts, raw API discovery, shell piping). Routes intent to the right transport, pre-flights the Notion-flavored Markdown spec and target data-source schema, and pins the empirical gotchas not surfaced by tool descriptions or `ntn --help`. Use whenever the user wants to read, fetch, search, create, update, query, or organize Notion content — pages, databases, data sources, views, comments, blocks, properties, schemas, wikis, docs — or upload a file to Notion, build a Notion Worker, or script Notion non-interactively. Triggers on Notion, Notion page, Notion database, Notion DB, Notion wiki, Notion doc, Notion comment, Notion view, Notion property, Notion schema, Notion file, Notion Worker, NOTION_API_TOKEN, ntn, ntn api.

Single-pass feature implementation using Explore → Code → Test. Ships focused changes at maximum speed, with a built-in circuit breaker that stops and recommends `/ultrapex` (or `/apex`) or `/forge` when the task turns out more complex than it looked. Use this whenever the user wants a quick win on a single, focused task — even when they don't say "oneshot" (e.g. "just", "quickly", "small change", "#42", or a GitHub issue URL for a small fix).

Turn a music brief into Suno v5.5-ready prompt artifacts — TRACK.md per generation, optional ALBUM.md for multi-track projects, optional ARTIST.md for artist-scoped identity. Artifact-emit-only — the user copy-pastes the prompt block into Suno's UI, listens, then iterates via `revise`. No API integration. Hardens the copyright contract — never lets artist names or copyrighted citations through into a prompt (legal exposure + Suno filters them anyway). Use when the user wants to make a song, produce a track or album, write a Suno prompt, or compose music with v5.5 — even when they don't say "Suno" by name.

Author, audit, or polish a project README — clarity, structure (Pattern A per-section collapse / Pattern B grouped / Pattern C per-entry), wording concision, anchor integrity. Reads the repo first, proposes diffs, applies on approval. Use when creating, restructuring, auditing, optimizing, or polishing a README for clarity, concision, or scannable structure — "write readme", "audit readme", "polish readme", "clarify readme", "optimize readme", "restructure readme", README clarity/wording — for long human-facing docs, skill libraries, npm SDK references, or CLI manuals.

Fixture for compound skills + npm tooling detection.

Fixture skill for repo-kind classifier tests.