Coding · Reviewed 2026-06-08

html-anything

FADING · 65/100

Purpose-built Claude Code and Codex skill that turns any file into a shareable single-file HTML — narrow scope, high execution quality, genuinely useful for the target audience.

Visit html-anything →

html-anything is an agent skill (skill.sh distribution) designed to run inside Claude Code and Codex. Its job is narrowly defined and executed well: take any input — WhatsApp exports, PDFs, CSV data, code files, Amazon Kindle exports, Spotify data, URLs — and produce a self-contained, offline-capable single-file HTML artifact with appropriate visual styling. The skill ships 60 source-aware parsing prompts and 17 style systems, which is a meaningful engineering investment for a focused tool. The skills.sh distribution model and the badge infrastructure show professional packaging for the agent-skill ecosystem. The limitation is scope: this is a transformation utility, not an agent. It produces artifacts; it does not act. For developers in the Claude Code/Codex ecosystem who regularly need to share structured data as web-readable outputs — without uploading to a third-party service or maintaining a server — html-anything fills a real gap. Outside that ecosystem, it has no distribution path.

Why FADING

FADING (65) because the tool is well-executed within its narrow scope, the 60-source and 17-style-system depth is credible, and the skills.sh distribution shows ecosystem maturity. Not STEADY because the scope is intentionally narrow (transformation-only, no autonomous action), the distribution is locked to Claude Code/Codex users, and adoption signals (star count) were not available in the pipeline run.

What it does well

What it fails at

Best for

  • Claude Code and Codex users who regularly need to share structured data as readable web pages
  • Developers who want to turn personal data exports (WhatsApp, Kindle highlights, Spotify history) into shareable HTML without uploading to third-party services
  • Agent workflows where the final deliverable needs to be a self-contained HTML artifact
  • Anyone who needs a consistent, styled output format across 60+ different input file types

Not recommended for

  • Developers not using Claude Code or Codex (no alternative invocation path documented)
  • Use cases requiring ongoing data updates — html-anything produces static snapshots, not live pages
  • Teams who need server-side generation (the skill runs in the agent session, not in a hosted service)
  • Non-technical users who cannot invoke agent skills

Compared to

Agent relevance

Behavioral-testable

Designed as a native skill for Claude Code and Codex. Invoked as part of an agent session — the agent calls the skill prompt which handles parsing and HTML generation. No API or CLI surface for external programmatic use.

Agent-friendly score: 7/10

Evidence

Public-surface checklist

scorecard.json · registry · methodology

Verdict by Hlido Editor · Method: public-surface-tier-2+editorial-narrative-v2 · Methodology version 2026.05 · Next review due 2026-09-08