A prompt to generate skills
distributed.camp
skill-generator
1copy and paste what's below here
2==================================================
3here's the structure for creating a claude skill. follow these rules and guidelines to a t.
4
5
6Claude Skills: Power User Guide
7What a Skill Actually Is
8A skill is a folder with a SKILL.md at its root, plus optional bundled resources. Claude reads the SKILL.md when triggered and follows its instructions like a mini system prompt for that task. There are no variables, loops, or iterators in the traditional programming sense — the "logic" is natural language instructions telling Claude what to do conditionally, sequentially, or branching.
9The Core Structure
10my-skill/
11├── SKILL.md ← required; triggers + instructions
12└── references/ ← optional deep docs loaded on demand
13└── assets/ ← optional templates, fonts, brand files
14└── scripts/ ← optional executable code Claude can run
15SKILL.md anatomy:
16markdown---
17name: my-skill
18description: [TRIGGER TEXT — this is what Claude pattern-matches against]
19---
20# Instructions
21...the actual directions...
22The Real "Variables" — The Description Field
23The description in the YAML frontmatter is the most important thing you'll write. Claude decides whether to load the skill based on this field alone. It's not code, but it functions like a matching rule.
24Weak description: "Helps with marketing copy"
25Strong description: "Write brand-aligned marketing copy, campaign briefs, launch announcements, ad copy, email sequences, or any customer-facing content. Use this skill whenever the user mentions campaigns, positioning, messaging, copy, launches, or asks to write anything that will be seen externally."
26Key tactics:
27Be pushy — skills undertrigger more than overtrigger
28Include synonyms and phrasings users actually type
29Mention roles ("when a marketer asks..." or "for finance team...")
30Include both the what and when
31"Control Flow" in Natural Language
32Since there's no code interpreter in SKILL.md, you use structured prose as logic:
33Branching (if/else):
34markdown## Step 1: Identify Document Type
35- If the user says "memo" → load `references/memo-format.md`
36- If the user says "board update" → load `references/board-format.md`
37- If unclear → ask: "Is this internal or external?"
38Sequential (loops via numbered steps):
39markdown1. Draft the section
402. Review against brand guidelines in `references/brand.md`
413. Revise if any guideline is violated
424. Repeat steps 1-3 for each section
43Conditionals:
44markdownIf the output is longer than 1 page, add an executive summary at the top.
45```
46This is the mental model: **you're writing directions for a very capable intern, not a compiler.**
47---
48### Progressive Disclosure — The 3-Level System
49This is the key architectural concept:
50| Level | What | Size | When loaded |
51|-------|------|------|-------------|
52| 1 | `name` + `description` | ~100 words | Always (every message) |
53| 2 | SKILL.md body | <500 lines ideal | When skill triggers |
54| 3 | `references/` files | Unlimited | Only when SKILL.md says to |
55**Practical implication:** Keep SKILL.md as a *router*, not an encyclopedia. Put deep content in reference files and tell Claude when to read them.
56---
57### Reference Files — How Many / How Long?
58- **SKILL.md body:** Under 500 lines. If you're going over, add a `references/` folder.
59- **Individual reference files:** 300+ lines → include a table of contents at the top so Claude can navigate efficiently.
60- **Number of files:** No hard limit, but organize by use case so Claude reads only what's relevant (e.g., `references/aws.md`, `references/gcp.md` — Claude reads only the one that matches the user's context).
61- **Assets (images, fonts, templates):** No size limit — Claude can use these without loading them into context.
62---
63### Startup Role Templates to Build
64Here's how this maps across your team:
65**Product**
66```
67references/
68 prd-template.md
69 user-story-format.md
70 feature-spec.md
71```
72Trigger on: "write a PRD", "spec this feature", "user stories for..."
73**Marketing**
74```
75references/
76 brand-voice.md
77 campaign-brief-template.md
78 email-sequences.md
79 launch-checklist.md
80```
81Trigger on: "campaign", "copy", "launch", "email sequence", "positioning"
82**Finance**
83```
84references/
85 financial-model-conventions.md
86 board-update-format.md
87 budget-narrative-template.md
88```
89Trigger on: "board deck", "budget", "financial summary", "runway analysis"
90**Ops/Recruiting**
91```
92references/
93 job-description-template.md
94 offer-letter-format.md
95 process-doc-template.md
96Power User Patterns
971. Chain skills via references
98SKILL.md says: "First read references/brand.md, then read references/channel-specific.md based on the platform."
992. Scripts for deterministic work
100Put Python/bash in scripts/ for things like file transforms, data formatting, or API calls. Claude runs them without loading them into context (efficient).
1013. Asset injection
102Put a Word/PPTX template in assets/ — the skill tells Claude to populate that template rather than build from scratch. Huge quality boost.
1034. Persona + constraints in SKILL.md
104markdown## Tone
105Write like a senior operator, not a consultant.
106Avoid jargon. Max 3 bullet points per section.
107Never use the word "leverage."
1085. Eval loop for iteration
109Write 5–10 test prompts, run them, review outputs. The skill-creator skill has a full eval harness if you want to go deep.
110Quickstart for a Non-Eng Skill
111Create my-skill/SKILL.md with a punchy description + 20–50 lines of instructions
112Add 1–2 references/ files with templates or style guides
113Test by typing a prompt that should trigger it and see if Claude reads the skill
114If it doesn't trigger — rewrite the description to be more specific and "pushy"
115
116
117i want you to create a skill for (super verbose explanation of a particular job to be done. consider not only the artifacts to be directly generated, but also any related artifacts to prove quality or success)
118if you think you can also generate meaningful reference files, do so.