docs — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited docs (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
hypertype is also an Agent Skill. The kit you can paste by hand is packaged so Claude and other AI coding tools reach for it when you ask for a headline, a hero, an editorial layout, a data table, or anything that should "look less generic."
It works because hypertype inlines. The assistants that write a single HTML file (Claude Design, Desktop, Cowork) can't run a build or npm install, and a <style> block plus a small <script> is the only thing that lands in all of them. So the skill tells the model to write exactly that.
git clone https://github.com/sceboucher/hypertype
cd hypertype && npm run install:skillThat installs the skill to ~/.claude/skills/hypertype/, where it registers as /hypertype. From then on the model will apply the kit when typography matters, or you can invoke it explicitly.
These do not load a skill from a repository, so paste the condensed instruction block instead. Copy skill/paste-block.md (about 870 tokens) into your project or system instructions. It carries the same guidance as the full skill, and both are generated from one source so they never drift.
The skill is more than a pointer to a library. It carries the judgment calls a person makes: pick a typeface instead of leaving the system default, build headings off a scale, switch on the OpenType features that suit the context, keep the line length readable, and use slab.js on display headlines. That's the difference between output that looks generated and output that looks set by hand.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.