← ClaudeAtlas

skill-authoring-stylelisted

Formatting and prose conventions for writing or editing a SKILL.md in this repo — horizontal rules, bold usage, line wrapping, spelling, negative instructions, and when to reach for a scripts/ helper. Use whenever creating a new skill or improving an existing one in this repository.
azborgonovo/ai-skills · ★ 0 · API & Backend · score 72
Install: claude install-skill azborgonovo/ai-skills
This repo holds Claude Code skills. SKILL.md files are read by Claude, not primarily by humans, so formatting should earn its tokens by helping the model parse and weight instructions — not by looking tidy. Apply these rules when creating a new skill or improving an existing one. ### Horizontal rules (`---`) Don't use `---` as section separators between steps or sections. The `##`/`###` Markdown headers already delimit sections. The **only** `---` that belong in a SKILL.md are the two YAML frontmatter delimiters at the very top (opening and closing the `name`/`description` block). ### Bold (`**...**`) Bold is a signal, not decoration — and signal weakens when it's everywhere. Use it purposefully: - **Keep it for labeled lead-ins** that act as inline sub-headers, e.g. `**Truncation check**:`, `**Why the script**:`. These help the model locate and weight a specific piece of guidance. - **Keep it for directive anchors** on things that change behavior, e.g. `**Never** approve`, `**Always** post through the script`, `**Critical**`. - **Drop it when it's purely cosmetic** — bolding a phrase mid-sentence for emphasis it doesn't need just dilutes the bold that does matter. When in doubt, ask: would the model behave differently if this weren't bold? If not, leave it plain. Don't strip bold wholesale either — losing the useful lead-ins and directives costs more in clarity than it saves in tokens. ### Line wrapping Don't hard-wrap prose in SKILL.md bodies. Write one line per pa