compact-markdown
Skill4 starsChanged 17 days agoMITView on GitHub
Writing style for markdown files. Strips filler, writes telegraphic and content-only — docs are for AI agents, not prose. Trigger whenever creating or editing ANY .md file (skills, rules, CLAUDE.md, feature specs, design notes, guides, etc.), for the new/edited content you produce. Do NOT trigger for README.md, and do NOT trigger for non-markdown files.
In this repository17 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- ask-firstSkill
- claude-project-rulesSkill
- humanizeSkill
- pr-reviewSkill
- empty-commitSkill
- github-issue-creatorSkill
- git-mergeSkill
- git-worktreeSkill
- handoff-sessionSkill
- html-landing-pageSkill
- offload-taskSkill
- readme-craftSkill
- skill-auditSkill
- study-notesSkill
- visualize-itSkill
- zoom-outSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 61
Remove anything that adds no information.
- Distinctive: line 63
Check that Markdown structure remains valid.
≈ “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
Compact Markdown
Markdown docs here exist for AI agents, not human reading pleasure. Write content only. No filler, no scaffolding.
Scope
- Apply to any
.mdbeing created or edited — exceptREADME.md(skip this skill entirely for README.md). - Apply only to content you write or edit. Do not rewrite existing surrounding prose you weren't asked to touch.
- Compress wording; keep markdown valid and navigable.
- Authoring a
SKILL.md? Readreferences/skill-authoring.mdfor skill-specific rules (frontmatter, triggering, what not to compact).
Core rule
Telegraphic. Drop articles, filler verbs, and pronouns where meaning survives.
- "You should run the command before testing" → "Run command before testing."
- "This function is responsible for parsing the config" → "Parses config."
- "In order to do X, you need to do Y" → "X requires Y."
- Imperative fragments are fine: "Fail → check trace." "No output → test passed."
Judgment overrides compression. Never compress if it loses important info. When in doubt, keep the words.
Strip
- Emojis — all of them.
- Badges — shields.io pills, image status badges (
). - Filler headers — empty
## Overview/## Introductionwhose body just restates the title. Collapse into actual content or open straight with facts. - Congratulatory / motivational lines — "Great!", "Congrats!", "That's it!", "Happy coding!", "Now you're ready to...".
- Hedging and throat-clearing — "It's worth noting that", "Basically", "Simply", "Of course", "As you can see".
- Restating the prompt or the section title in the first sentence.
- Summary / Conclusion sections that only restate the doc. Keep summaries that capture decisions, risks, or takeaways.
- Table of contents for docs short enough to scan.
- Bold on every other phrase — everything emphasized = nothing emphasized.
- Horizontal rules (
---) between every section.
Keep
- Code blocks, inline code, CLI commands — verbatim, never paraphrased.
- Examples and sample snippets — they carry the most signal per token. Drop only a duplicate/low-value example, never a useful one.
- Tables — keep structured, do not flatten into prose.
- Heading hierarchy — keep for navigation. Just cut filler headers.
- Any concrete detail: paths, flags, names, values, edge cases, gotchas.
Shape
- Reference data (commands, paths, options, values) → lists and tables, one fact per line.
- Behavioral rules → keep the rule and its reason together in one line or sentence; a bullet that drops the "because" loses the rule's scope.
→for cause/result, conditions, flows.- Lead with the rule, then the why (only if non-obvious), then example.
Self-check
Before finishing:
- Remove anything that adds no information.
- Restore any detail whose removal changes meaning, scope, or conditions.
- Check that Markdown structure remains valid.
Example
Verbose:
## Introduction
This section will walk you through the process of setting up the
authentication flow. It's worth noting that you'll need to make sure
the environment variables are configured correctly before you begin.
Great! Once that's done, you can run the setup command, `npm run setup:auth`. 🎉Compact:
## Auth setup
Configure env vars first, then run:
```bash
npm run setup:auth
```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.

