povvo/claudikins-github-readme-for-perfectionists/pen-wielding
Claude Code and agent skill by povvo/claudikins-github-readme-for-perfectionists
- community
Checking that the SKILL.md file is still in the repository.
Use when writing the final README after all phases complete. Synthesises outputs from deep-dive, crystal-ball, brain-jam, and think-tank. Applies Anti-Slop style guide.
Install
npx vskill@latest install povvo/claudikins-github-readme-for-perfectionists/pen-wieldingSKILL.md
Pen-Wielding Protocol
Goal: Write a production-ready GitHub README that feels authentically human.
Step 1: Ingest Context
Read the outputs from the preparatory skills:
- Facts:
deep-divereport (Stack, friction points, install complexity) - Future:
crystal-ballanalysis (Roadmap table, technical debt) - Voice:
brain-jamdecision (The chosen angle: Deep Tech vs Pragmatic) - Plan:
think-tankfindings (Selected badges, sections, and visual strategy)
Do not proceed until all four reports are available.
Step 2: Load Standards
Read the governance files to establish the form:
references/style-guide.md(Anti-Slop rules, Banned Words, Humanisation Patterns, Reusable Phrases)references/structural-templates.md(Section templates, Results/Metrics tables, Do/Don't patterns)references/visual-engineering.md(Density rules, diagrams, ASCII visualisations)references/anti-patterns.md(Final checklist, Writing/Research anti-patterns)
Note: The reference files include patterns integrated from the influencer-skills-synthesis research (archived at ../grfp/reference/influencer-skills-synthesis.md).
Step 3: Writing Rules (Always Enforced)
Style
- Clear and concise, active voice
- Important information first
- No M dash (—) EVER - use hyphens or restructure
- Technical only - no marketing fluff
- No paragraph > 4 sentences
The "Anti-Slop" Filter
- Check every sentence against the Banned Words list (No "Delve", "Seamless", "Unleash")
- Never use "The Problem / The Solution" framing - it's lazy and cliched
- Enforce Patterns:
- Use "The Hook" (Pattern A) in the _Description_
- Use "The Hammer" (Pattern B) in the _Features_ list
- Use "The Trust Builder" (Pattern C) in _Usage/Configuration_
Spelling Consistency
Pick one and stick to it:
- British: colour, optimise, analyse, behaviour, centre, finalising
- American: color, optimize, analyze, behavior, center, finalizing
Step 4: Section Templates
Write sections in this order:
1. Hero Section
2. Description
3. Installation
4. Quick Start
5. Usage (with collapsible advanced)
6. Configuration (table format)
Step 5: Visual Engineering
When to Use What
| Visual Type | Use Case | Tool | |
|---|---|---|---|
| Terminal GIF | CLI tools, workflows | VHS | |
| Diagram | Architecture, data flow | Mermaid | |
| Screenshot | UI apps, dashboards | With alt text |
Visual Density
Target: 1 visual per 300 words. If exceeded, add code block, diagram, or break into subsections.
Mermaid Diagrams (GitHub-native)
Visual Placeholders
Insert specific placeholders for visuals identified in think-tank:
Accessibility Requirements
- Alt text: Always descriptive, never empty
- Bad:
 - Good:

Step 6: Anti-Patterns (Never Do)
| Anti-Pattern | Problem | Fix | |
|---|---|---|---|
| Wall of text | Unreadable | Max 4 sentences per paragraph | |
| Badge overload | Cluttered | Max 5-7 essential badges | |
| Generic headers | Poor SEO | Semantic: "High-Performance Parsing" not "Features" | |
| Missing prerequisites | Broken installs | Explicit versions: "Node.js 18+" | |
| No output shown | Confusing | Always show expected result | |
| M dash (—) | Rendering issues | Use hyphens or restructure | |
| Buried quick start | User abandonment | Within first screen view | |
| Copy-paste friction | User errors | Working defaults, note what to change | |
| Click here links | Poor accessibility | Descriptive link text | |
| H1 just project name | Poor SEO | Include core function | |
| "The Problem / The Solution" | Cliched, lazy | Show value directly |
Step 7: Quality Metrics
| Metric | Target | Action if Violated | |
|---|---|---|---|
| Flesch-Kincaid | Grade 8-10 | Simplify sentences | |
| Time to Joy | <=3 commands | Add Docker/Makefile | |
| Visual density | 1 per 300 words | Add code block/diagram | |
| Badge count | 5-7 | Curate to essentials | |
| Quick start visibility | < 30 seconds | Move above the fold | |
| Delve Index | 0 | Remove banned words |
Step 8: Verification Against Previous Steps
Before finalising, cross-check against each preparatory step:
Deep Dive Verification
- [ ] Tech stack mentioned matches deep-dive findings
- [ ] Prerequisites include all system dependencies identified
- [ ] Entry points documented match identified entry points
- [ ] Installation steps work for the detected ecosystem
Crystal Ball Verification
- [ ] Roadmap section included (if opportunities identified)
- [ ] Future features mentioned align with crystal-ball analysis
- [ ] Technical debt items acknowledged where relevant
- [ ] Version/maturity signals match project state
Brain-Jam Verification
- [ ] Tone matches the chosen angle (Deep Tech vs Pragmatic vs etc.)
- [ ] Voice is consistent throughout
- [ ] Marketing angle applied to hero section and description
- [ ] Target audience language used appropriately
Think-Tank Verification
- [ ] Badge selection matches think-tank recommendations
- [ ] Section order follows the researched blueprint
- [ ] Visual strategy implemented as planned
- [ ] Patterns borrowed from exemplars applied correctly
- [ ] Anti-patterns from research avoided
Step 9: The Final Audit
Before outputting, review against references/anti-patterns.md:
- [ ] Are paragraphs < 4 sentences?
- [ ] Are M-dashes (—) removed?
- [ ] Is the "Delve" index 0?
- [ ] Are headers semantic (e.g., "Zero-Config Build" vs "Features")?
- [ ] Does the Quick Start work in < 30 seconds?
- [ ] No "The Problem / The Solution" framing used?
Output
Present the final README in a single, copy-pasteable Markdown block.
Ask: "README complete. Want to refine anything, or ship it?"
- If refine -> make edits
- If ship -> write to README.md, commit, done
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/pen-wielding/SKILL.md
Popularity
- Repository stars
- 3
- Repository forks
- 1
- 7-day trend score
- 7
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/povvo/claudikins-github-readme-for-perfectionists/pen-wielding)