npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@zakelfassi/htmlify

v1.0.0

Published

Turn agent answers into self-contained HTML artifacts and presentation decks — operator briefs, review packets, incident timelines, decision briefs, and talks. Zero dependencies, zero build, one file.

Downloads

319

Readme

Coding agents answer in walls of markdown. When the answer has shape — a comparison, an architecture, a timeline, a review — that wall flattens it. htmlify is a skill family that makes agents ship designed, self-contained HTML instead: operator briefs, PR review packets, incident timelines, decision briefs, explainers, interactive boards, and full presentation decks with speaker notes. Zero dependencies, zero build step, one auditable file.

See it live: the gallery — every artifact there was generated by these skills, about this repository.

What's in the box

| Piece | What it is | | --- | --- | | skills/htmlify | Document skill — 10 modes: operator-brief, build-plan, implementation-map, pr-review-packet, release-brief, incident-report, decision-brief, status-report, explainer, prototype/editor | | skills/deckify | Deck skill — 6 modes: talk-deck, workshop-deck, essay-deck, demo-deck, launch-deck, teaching-guide; speaker notes, run-of-show, downloadable guide/PDF | | htmlify-answer CLI | Pipe any text into a designed artifact; validate any artifact against the rich/app/deck safety profiles | | Pi / OMP extension | /htmlify commands, render modes, browser annotation layer | | Claude Code plugin | One-command install of both skills, optional long-answer Stop hook | | Hardcopy | The design system every artifact ships in — engineering-plate language: warm paper, ink hairlines, serif display, mono metadata, one signal-orange accent |

Install

| Agent | Install | | --- | --- | | Claude Code | /plugin marketplace add zakelfassi/htmlify then /plugin install htmlify@htmlify | | Codex | git clone https://github.com/zakelfassi/htmlify.git ~/.htmlify && ln -sfn ~/.htmlify/skills/htmlify ~/.codex/skills/htmlify && ln -sfn ~/.htmlify/skills/deckify ~/.codex/skills/deckify | | Cursor / Windsurf | Clone the repo, point a project rule at skills/htmlify/SKILL.md / skills/deckify/SKILL.md | | Aider / anything | printf '%s' "$ANSWER" \| npx -y @zakelfassi/htmlify htmlify-answer --title "Review" | | Pi / Oh-My-Pi | pi install npm:@zakelfassi/htmlify |

Per-agent recipes, project rules, and hook setup: agent-integrations.md.

How it works

  1. Evidence first. The skill reads the repo, git state, PRs, CI, and logs before designing anything; unverifiable claims are stamped needs verification.
  2. One HTML file. Inline CSS+JS in the Hardcopy design language. No CDNs, no fonts, no analytics, no build.
  3. Prove it. The bundled validator gates the output — then it opens in your browser.
npx -y @zakelfassi/htmlify htmlify-answer --validate artifact.html --profile auto

| Profile | Rules | | --- | --- | | rich | No scripts at all — for model-generated documents | | app | Inline scripts allowed (editors, boards); external scripts, on*= handlers, javascript: URLs still banned | | deck | app plus the deck contract: ≥2 slides, keyboard nav, speaker notes on substantive slides, print CSS | | auto | Detected per file from the document shape |

Exit codes: 0 valid · 1 validation errors · 2 usage/IO. Add --format json for agent consumption.

Artifact modes

| Mode | Use when | Live example | | --- | --- | --- | | operator-brief | What happened, what's next, risks, attention | operator-brief | | pr-review-packet | Motivation, diff tour, reviewer checklist | PR #1 packet | | incident-report | Impact, timeline, root cause, follow-ups | capture bug | | decision-brief | Options, tradeoffs, the call | monorepo decision | | implementation-map | Modules, data flow, hot path | runtime map | | explainer | Concepts, comparisons, glossary, FAQ | HTML vs markdown | | prototype / editor | Interactive triage, tuning, ordering — with export | launch board | | talk-deck (deckify) | Talks with speaker notes + run-of-show | launch talk | | workshop-deck (deckify) | Teaching with exercises + printable guide | skill workshop |

Plus build-plan, release-brief, status-report, and deckify's essay-deck, demo-deck, launch-deck, teaching-guide.

The Pi / OMP runtime (optional)

The npm package doubles as a Pi / Oh-My-Pi extension that captures long answers and exports them on demand:

| Command | Result | | --- | --- | | /htmlify | Quick local HTML of the last long answer (Hardcopy-styled, outline rail, annotation layer) | | /htmlify choose | Render-mode chooser — local, pi (current model second pass), gemini (Gemini CLI, falls back to local) | | /htmlify-comments <json> | Import browser review comments back to the agent as a structured prompt | | /htmlify-version | Show loaded version |

Legacy aliases (/html-last, /html-comments, /html-last-version) keep working. Exports include a trusted annotation layer: select text in the browser, comment, copy Markdown for the agent, or download a JSON bundle. Long answers stay visible in the terminal — export is always explicit.

For agents with hooks, the bundled Claude Code Stop hook archives answers past a threshold (HTMLIFY_MIN_CHARS, default 2500) to HTMLIFY_EXPORT_ROOT — opt-in, see agent-integrations.md.

Migrating from 0.x

1.0 restructures the repo into a skill family. Breaking: the root SKILL.md and references/ moved to skills/htmlify/; clones installed as a skill directory at the repo root must re-install:

ln -sfn /path/to/htmlify/skills/htmlify ~/.codex/skills/htmlify   # and skills/deckify

Unchanged: the htmlify-answer CLI flags and stdin behavior (--validate is additive), the hook path hooks/claude-code-stop-htmlify.js (existing settings.json entries keep working), the Pi/OMP entry points, all /htmlify commands, and the env vars (including the legacy PI_HTML_LONG_ANSWER_* aliases).

Development

corepack enable && pnpm install
pnpm test        # node --test, 35 tests
pnpm lint        # biome
pnpm typecheck   # strict tsc over JSDoc types
node bin/htmlify-answer.js --validate examples/htmlify/*.html examples/deckify/*.html --profile auto

The runtime is dependency-free CommonJS under src/ with strict JSDoc type checking — what ships is what you read. Releases are cut by release-please from conventional commits, published to npm with provenance. See CONTRIBUTING.md.

Trust and security

Extensions and hooks run with your user permissions — install from sources you trust and pin a ref when you need reproducibility. Model-generated HTML is treated as untrusted until validated: scripts (in the rich profile), event-handler attributes, javascript: URLs, external assets/CSS, meta refresh, and oversized output are rejected, with fallback to the local renderer. Interactive profiles still ban every external-execution vector. Found a way around the validator? That's a security report we want: SECURITY.md.

License

Apache-2.0 © Zak El Fassi