style-doc
Skill2 starsChanged 4 months agoMITView on GitHub
Create, restyle, or present reviewed standalone HTML technical reports in a cohesive editorial system. Use for new reports, visual restyles, or corrected copies with tracked changes; report-producing skills may invoke it for standalone HTML presentation.
In this repository12 other files it gives its agents.
- adversarial-code-reviewSkill
- code-review-loopSkill
- effect-loggingSkill
- finalizeSkill
- fly-optimizationSkill
- formattingSkill
- product-design-evolveSkill
- product-design-extendSkill
- product-design-propagateSkill
- simplifySkill
- test-value-auditSkill
- typescript-structureSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 4 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 16
If branches overlap, use reviewed whenever corrections must remain visible.
- Very specific: line 20
Use one valid standalone HTML file with one main, one h1, semantic sections, and a descriptive title.
- Very specific: line 21
Keep a white canvas, zinc typography, blue review accents, thin dividers, restrained surfaces, and a max-w-4xl reading measure.
- Distinctive: line 22
Use Inter when available with a system sans-serif fallback.
- Distinctive: line 24
Keep claims evidence-backed, uncertainty explicit, headings in sentence case, and editorial changes distinguishable without relying on …
- Distinctive: line 28
Read every reference whose condition applies before writing the report:
- Distinctive: line 46
Before delivery, verify every applicable condition:
- Very specific: line 48
Coverage: Every source section, finding, metric, citation, and caveat is present, or its intentional omission is documented.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Editorial Technical Reports
Produce a standalone technical document, not a dashboard or marketing page. Preserve the report's evidence-first character while making its hierarchy, caveats, and actions easy to scan.
Route the Request
- Create: Turn source findings or a brief into a new standalone HTML report. Establish information hierarchy without inventing evidence.
- Restyle: Apply this system to an existing report. Preserve its claims, ordering, citations, and caveats unless the user also requests editorial changes.
- Reviewed: Present a corrected or adversarially reviewed copy. Preserve traceability between material source claims and corrections.
If branches overlap, use reviewed whenever corrections must remain visible; otherwise use restyle for existing content and create for new content.
Core Invariants
- Use one valid standalone HTML file with one
main, oneh1, semantic sections, and a descriptivetitle. - Keep a white canvas, zinc typography, blue review accents, thin dividers, restrained surfaces, and a
max-w-4xlreading measure. - Use Inter when available with a system sans-serif fallback; use monospace for paths, identifiers, and code.
- Let evidence determine the component: prose for reasoning, definition lists for metadata and metrics, semantic tables for comparison, and reviewer notes for caveats.
- Keep claims evidence-backed, uncertainty explicit, headings in sentence case, and editorial changes distinguishable without relying on color.
Load References by Context
Read every reference whose condition applies before writing the report:
| Condition | Required reference |
| Creating a file or changing its fonts, CSS delivery, or runtime dependencies | Read references/standalone-shell.md before choosing the networked or self-contained shell. |
| Creating or restyling report hierarchy, spacing, typography, metadata, metrics, or reviewer notes | Read references/report-anatomy.md before composing sections. |
| The report contains any evidence, inventory, risk, route, or roadmap table | Read references/evidence-tables.md before emitting the first table. |
The branch is reviewed, or source claims must remain visibly corrected | Read references/tracked-changes.md before marking any correction. |
The create branch needs a baseline composition | Read references/neutral-example.md after the shell and anatomy references, then replace all sample content. |
Application Process
- Select
create,restyle, orreviewed, and inventory the source sections, findings, metrics, citations, and caveats. Complete when every source item has a destination or an intentional omission requested by the user. - Load each context-triggered reference above. Complete when every applicable reference has informed the chosen structure and components.
- Build or update the standalone HTML while preserving source meaning and applying the editorial system. Complete when the document opens as one coherent report and all source items are accounted for.
Completion Gate
Before delivery, verify every applicable condition:
- Coverage: Every source section, finding, metric, citation, and caveat is present, or its intentional omission is documented; the
restylebranch preserves meaning, and thereviewedbranch keeps material corrections traceable. - Document integrity: The file has a doctype, language, UTF-8 charset, viewport, descriptive title, one
main, oneh1, globally unique IDs, valid heading order, and no placeholder or sample content. - Dependency integrity: The chosen network mode is explicit and consistent; networked assets load successfully, or the self-contained file renders with no runtime network dependency.
- Semantics and access: Metadata and metrics use definition lists; data comparisons use real tables with captions and scoped headers; links, focus states, labels, and editorial marks remain understandable without color.
- Visual system: Typography, measure, spacing, dividers, notes, tables, and status labels form one restrained editorial hierarchy; editorial-change colors stay separate from operational risk/action colors.
- Responsive behavior: At 390px and 1440px viewports, body text is legible, headings wrap cleanly, content stays within the page, and each wide table scrolls inside its labeled region without causing page-level horizontal overflow.
- Evidence quality: Claims use the strongest available evidence, unverified statements are qualified, adversarial claims include source locations, and reviewer notes contain a concrete caveat, source, or sequencing constraint.
Deliver only when every applicable condition passes.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

