Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Guideline Writing

A craft skill for writing and reviewing the instruction layer itself — CLAUDE.md, .claude guides, policies and plan files — with operative rules, force tiers and bloat control.

Content & WritingIntermediate★ 142⑂ 24AI score 8/10Last updated: Sep 13, 2026

What it does

  • Splits every sentence into operative rule vs recital, keeping only obligations in the body and exiling deliberation narrative to an append-only log.
  • Forces each rule onto exactly one MUST / SHOULD / MAY tier with visible wording and a matching verification surface (lint gate, advisory hook, or none).
  • Applies a 5-step bloat self-check (absorb before adding, no copied tables, no new risk sections, re-review at ≥50 changed lines, read whole files) plus 4 red flags.
  • Defines a blind review protocol: no author intent, per-hunk substantive/invariant verdicts, a strict defect form (criterion id · location · issue · fix direction · severity), and a self-containment probe.
  • Ships a 15-row antipattern table with detection method and resolution for each.

Who it's for

  • Anyone maintaining CLAUDE.md, agent instructions, runbooks or internal policy docs.
  • Teams whose guidelines keep growing while compliance keeps dropping.
  • Reviewers who need an objective rubric for instruction-file diffs.

Examples

  1. Collapse a paragraph explaining why hook logic was consolidated into one enforceable line, and move the comparison story to the log.
  2. Rewrite "it is usually best to run the linter before handing off" as an unambiguous MUST with a stated stopping condition.
  3. Instead of adding a new "Drift Requirements" section, add one row to the existing drift table and cross-reference it.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/alfadur7/llm-wiki-newsroom/HEAD/.claude/skills/guideline-writing/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the .claude/skills/guideline-writing folder from the GitHub repo alfadur7/llm-wiki-newsroom into my ~/.claude/skills/guideline-writing/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/alfadur7/llm-wiki-newsroom /tmp/llm-wiki-newsroom && mkdir -p ~/.claude/skills && cp -r /tmp/llm-wiki-newsroom/.claude/skills/guideline-writing ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repo: git clone https://github.com/alfadur7/llm-wiki-newsroom /tmp/llm-wiki-newsroom
  3. Create the skills directory: mkdir -p ~/.claude/skills
  4. Copy the skill: cp -r /tmp/llm-wiki-newsroom/.claude/skills/guideline-writing ~/.claude/skills/
  5. Verify SKILL.md, criteria.json and checks.py exist under ~/.claude/skills/guideline-writing/.
  6. Restart Claude Code and try: "Review my CLAUDE.md against the guideline-writing criteria."
  7. Remap or ignore origin-repo paths such as log.md, tools/lint.py and .claude/agents/editor-in-chief.md so they match your own project layout.
View source on GitHub ↗License: MIT