メニュー
Scaffold a complete extraction pipeline for a new domain — Pydantic model, field definitions, example script, and tests. Use when building a new extraction use case like medical records, invoices, or reviews.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
SOC 職業分類に基づく
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/jhd3197/Prompture --skill scaffold-extractionコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
Update LLM model pricing in Prompture. Pricing resolves through a pluggable source registry — by default local KB JSON files (primary, curated) then models.dev (fallback). Use when model prices change, new models launch, models.dev data is stale, or you need to add a custom pricing source.
Scaffold a new LLM provider driver for Prompture. Creates sync + async driver classes, registers them in the driver registry, adds settings, env template, setup.py extras, package exports, discovery integration, and models.dev pricing. Use when adding support for a new LLM provider.
Create a new Prompture usage example script. Follows project conventions for file naming, section structure, docstrings, and output formatting. Use when demonstrating extraction use cases or provider integrations.
| name | scaffold-extraction |
| description | Scaffold a complete extraction pipeline for a new domain — Pydantic model, field definitions, example script, and tests. Use when building a new extraction use case like medical records, invoices, or reviews. |
| metadata | {"author":"prompture","version":"1.0"} |
Builds the full pipeline for a new extraction domain: model, fields, example, and tests.
Ask the user:
ollama/gpt-oss:20b| Scenario | Method |
|---|---|
| Simple model, < 8 fields | extract_with_model (1 LLM call) |
| Complex model, 8+ fields | stepwise_extract_with_model (N calls, more accurate) |
| No Pydantic model, raw schema | extract_and_jsonify |
| Structured input (CSV, JSON) | extract_from_data (TOON, 45-60% token savings) |
| DataFrame input | extract_from_pandas |
| Non-JSON output | render_output |
from pydantic import BaseModel, Field
class InvoiceData(BaseModel):
vendor_name: str = Field(description="Company that issued the invoice")
total_amount: float = Field(description="Total amount due")
currency: str = Field(default="USD", description="ISO 4217 currency code")
Rules:
Field(description=...) on every field — these become LLM instructions"", 0, 0.0, []Optional[T] = None only when a field genuinely might not existRegister in prompture/field_definitions.py only if fields are general-purpose. Use the add-field skill.
Create examples/{domain}_extraction_example.py. Use the add-example skill.
Unit test for the Pydantic model + integration test for end-to-end extraction. Use the add-test skill.
from prompture import extract_with_model
result = extract_with_model(
model_cls=InvoiceData,
text=invoice_text,
model_name="ollama/gpt-oss:20b",
instruction_template="Extract all invoice details from the following document:",
)
invoice = result["model"] # Pydantic instance
usage = result["usage"] # Token counts and cost