Skip to content
Skills Library
AI & agents · Writing

structure-prompt

/structure-prompt

Turn a rough request into a structured, production-grade prompt. Classifies which prompting dimensions apply (product verification, structured high-stakes detail, constraints-upfront, scannable structure, search priority, internal-first sourcing), fills a prompt template, and returns a copyable prompt plus a note on what was applied or omitted. Use for prompt architect, structure a prompt, harden a prompt, or turn an ask into a production prompt.

Licence
MIT
Last repository change
23.10.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
4
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 23.10.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Turn a rough request into a structured, production-grade prompt. Classifies which prompting dimensions apply (product verification, structured high-stakes detail, constraints-upfront, scannable structure, search priority, internal-first sourcing), fills a prompt template, and returns a copyable prompt plus a note on what was applied or omitted. Use for prompt architect, structure a prompt, harden a prompt, or turn an ask into a production prompt.

Original author description

Turn a rough request into a structured, production-grade prompt. Classifies which prompting dimensions apply (product verification, structured high-stakes detail, constraints-upfront, scannable structure, search priority, internal-first sourcing), fills a prompt template, and returns a copyable prompt plus a note on what was applied or omitted. Use for prompt architect, structure a prompt, harden a prompt, or turn an ask into a production prompt.

How it works

  1. Role
  2. When to use
  3. The six dimensions
  4. Workflow
  5. Prompt template
  6. Principles
  7. Safety
  8. Distinction from other commands
  9. Guardrails
Original SKILL.md headings · 109 lines, 788 words
SKILL.md preview
---
name: structure-prompt
description: >-
  Turn a rough request into a structured, production-grade prompt. Classifies
  which prompting dimensions apply (product verification, structured high-stakes
  detail, constraints-upfront, scannable structure, search priority, internal-first
  sourcing), fills a prompt template, and returns a copyable prompt plus a note on
  what was applied or omitted. Use for prompt architect, structure a prompt, harden
  a prompt, or turn an ask into a production prompt.
user-invocable: false
---
# Structure prompt

## Role

You act as a **prompt architect**. The user gives a rough request; you return a structured, copyable prompt that another agent or model can run, plus a short note on which dimensions you applied or omitted and why. You **produce a prompt**; you do not execute the underlying task in the same turn.

Keep the output **model-agnostic**: do not hardcode a specific model name or assume a particular vendor's leaked system prompt is canonical.

## When to use

Use when a user has an ask they want turned into a reliable prompt, especially high-stakes, ambiguous, or research-heavy work where output format and sourcing matter. For debugging an existing prompt with a tiny eval suite, use `prompt-eval-debug`. For designing a six-part delegation Goal before autonomous work, use `define-agent-goal`.

## The six dimensions

Apply only the ones that fit the request. Do not bolt on all six by default.

| Dimension | Trigger | What to add |
|-----------|---------|-------------|
| **Product verification** | Asks about a tool/product's current capabilities, pricing, or APIs | A clause to verify against current docs/support before answering; flag that product knowledge may be stale |
| **Structured high-stakes** | Output quality is critical | Concrete detail, positive **and** negative examples, step-by-step reasoning, XML tags, explicit length/format constraints |
| **Constraints upfront** | Request is ambiguous | Audience, format, scope, sources, success criteria, allowed tools, forbidden moves, review requirements |
| **Scannable structure** | Reader needs to skim or extract | Ask for explicit headings and bullets (otherwise prefer prose) |
| **Search priority** | Needs current/changing information | Declare source order, e.g. "primary docs first, then primary sources, then high-quality secondary coverage" |
| **Internal-first sourcing** | Touches company/org/personal data | Internal sources first, public second, synthesis last |

## Workflow

Run phases in order.

### Phase 0: Intake

1. Capture the **rough request** and the **target** (which model/agent will run it; chat vs production).
2. If the request is missing, ask one focused question before proceeding.
3. Produce a one-sentence **summary** of what the prompt must accomplish.

### Phase 1: Classify

Decide which of the six dimensions apply. State the verdict briefly (applied vs not applicable). Match effort to stakes: a trivial ask gets a light prompt, not the full template.

### Phase 2: Fill the template

Compose the structured prompt using only the applicable dimensions. Use the template below as a scaffold; drop sections that do not apply.

### Phase 3: Deliver

Post in chat:

1. The structured prompt in a single copyable code block.
2. A short **applied/omitted** note (which dimensions you used and which you skipped, with one-line reasons).

## Prompt template

Adapt and prune to fit the request:

```text
<role>Who the model is and who the audience is.</role>

<task>The specific outcome, with concrete detail.</task>

<constraints>
- Scope: what is in and out
- Format: structure of the output (headings/bullets vs prose)
- Success criteria: what "done" means
- Allowed tools / sources
- Forbidden moves
- Review requirements
</constraints>

<sources>
Search/source priority when current info is needed, e.g.
primary docs first, then primary sources, then secondary coverage.
For company data: internal sources first, public second, synthesis last.
</sources>

<examples>
Positive example: ...
Negative example (what to avoid): ...
</examples>

<reasoning>Think step by step before answering.</reasoning>

<output>Length and format constraints; the exact shape of the answer.</output>
```

## Principles

- **Proportionality**: simple asks get a light touch; do not over-engineer with all six dimensions.
- **Produce, don't run**: output a prompt; do not perform the task it describes.
- **Model-agnostic**: no hardcoded model names; no claim that a leaked system prompt is canonical.
- **Verify product facts**: when the prompt asks about a tool's current behavior, instruct it to check current docs rather than assert from memory.

## Safety

- Redact secrets, tokens, credentials, PII, and PHI in both examples and the produced prompt.
- Do not invent product capabilities, pricing, or model names.
- Do not use employer, product, or internal repository names in examples. Use generic names like `service-a`.

## Distinction from other commands

- **`prompt-eval-debug`**: debugs an existing prompt with a tiny eval suite and smallest next change, not prompt construction.
- **`instruction-ablation`**: thins an instruction stack via bare-run evidence and repeated failures. This skill **composes** a new structured prompt from a rough ask.
- **`define-agent-goal`**: six-part delegation Goal (outcome, verification, constraints, boundaries, iteration, stopping) before autonomous work, not a runnable prompt.

## Guardrails

- Match effort to stakes: keep simple asks proportional and do not bolt on all six dimensions.
- Produce the prompt and stop; do not execute the underlying task it describes.
- Verify product facts against current docs and redact secrets; do not invent capabilities, pricing, or model names.
Files in this package (4)
  • SKILL.md 6 KB
  • eval/README.md 1 KB
  • eval/cases.md 3 KB
  • eval/fixtures.yaml 1 KB

Versions

The database has no other skills named structure-prompt.

Identical content appears in 1 other GitHub folders. We show one copy with the highest score.

Database snapshot: 13.11.2026 · package version a16ff68036b1 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100ai-fluency-prompt-architect
    Turns rough intent into a copy-paste, production-grade AI prompt using the AI Fluency 4D framework (Delegation, Description, Discernment, Diligence) plus six durable prompting techniques. WHEN: \"write me a prompt\", \"make this prompt better\", \"improve my prompt\", \"prompt…
  • 100factory-prompting
    Structured prompting with XML tags. A small named vocabulary (`instructions`, `context`, `input`, `output_format`, `examples`, `constraints`, `role`, `thinking`) that turns a prose ask into something the model can parse without ambiguity. Read whenever you're authoring a prompt…
  • 100write-a-prompt
    Creates a copy-ready prompt from a rough request, notes, source material, or the current conversation using OpenAI's prompting guidance. Use when the user invokes $write-a-prompt or /write-a-prompt, types a common misspelling such as /write-a-promopt or /wite-a-prompt, asks to…
  • 100prompt-engineering-patterns
    This skill should be used when the user asks to "optimize a prompt", "improve prompt performance", "design a prompt template", "write better prompts", "debug prompt issues", "use chain-of-thought", "structured prompting", "few-shot prompting", or wants to apply advanced prompt…
  • 100prompt-engineering-patterns
    This skill should be used when the user asks to "optimize a prompt", "improve prompt performance", "design a prompt template", "write better prompts", "debug prompt issues", "use chain-of-thought", "structured prompting", "few-shot prompting", or wants to apply advanced prompt…
  • 100prompt-improver
    Turns a rough request or weak prompt into a strong one. Use when the user asks to write, improve, review or debug a prompt, system prompt, custom instructions, an agent's instructions, or a reusable template; or when a prompt keeps producing poor results. Checks the six elements…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.