Skip to content

Instantly share code, notes, and snippets.

View Tuff-Madman's full-sized avatar

Karcus with a C Tuff-Madman

View GitHub Profile
@Tuff-Madman
Tuff-Madman / README.md
Last active July 25, 2025 09:14
titel format and structure.md

titel format and structure.

FORMATTING RULES (Porta TITLE Formatter)

MEASUREMENTS

Rule M1 (Spacing): Always use a space between number and unit.

  • ✅ Correct: '250 cm', '75 kg', '1 200 mm', '2 l'
  • ❌ Incorrect: '250cm', '75kg', '1200mm', '2l'
@Tuff-Madman
Tuff-Madman / extract.csv
Last active July 25, 2025 09:04
titles
index sample_product_title structure
1 Schwingstuhl MARC 4er Set grau/chromfarbig Produkttyp + Eigenname + Stückzahl + Farbe/Finish
2 Beschlagset für Wandmontage verzinkt/ weiß Produkttyp + Zweck + Finish/Farbe
3 Schwingstuhl STEVE 2er Set 43 x 48 cm Stoffbezug grau Produkttyp + Eigenname + Stückzahl + Maße + Material + Farbe
4 Stuhl ZERO grün/schwarz Produkttyp + Eigenname + Farbe
5 Lowboard JOHN 220 x 44 cm Weiß/ Eiche Artisan Oak Produkttyp + Eigenname + Maße + Farbe/Material
6 Highboard MARBELLA 144 x 134 cm weiß/ braun Produkttyp + Eigenname + Maße + Farbe
7 CASAVANTI Schwingstuhl ALMADA II dunkelgrau/chromfarbig Marke + Produkttyp + Eigenname + Farbe
8 INTERhome Lowboard KAMERUN 160 x 50 cm Mango massiv braun Marke + Produkttyp + Eigenname + Maße + Material + Farbe
9 CASAVANTI Armlehnstuhl VALLETTA 2er Set Rostbraun Marke + Produkttyp + Eigenname + Stückzahl + Farbe
@Tuff-Madman
Tuff-Madman / roo_workflow.md
Created July 8, 2025 21:28 — forked from livecodelife/roo_workflow.md
Roo Code Setup and Workflow for the best $0 development

Roo Code Workflow: An Advanced LLM-Powered Development Setup

This gist outlines a highly effective and cost-optimized workflow for software development using Roo Code, leveraging a multi-model approach and a custom "Think" mode for enhanced reasoning and token efficiency. This setup has been successfully used to build complex applications, such as Baccarat game simulations with betting strategy analysis.


Core Components & Model Allocation

The power of this setup lies in strategically assigning different Large Language Models (LLMs) to specialized "modes" within Roo Code, optimizing for performance, cost, and specific task requirements.

@Tuff-Madman
Tuff-Madman / README.md
Last active July 3, 2025 12:39
CUSTUM INSTRUCTION CLINE

CUSTUM INSTRUCTION CLINE

Core Operational Directive: The Baby Steps™ Methodology

Preamble: You are to operate under a single, overriding principle: the Baby Steps™ Methodology. Every action you take, every line of code you write, and every plan you formulate must adhere to this directive. Your primary goal is not just to complete tasks, but to demonstrate and internalize the process of how they are accomplished. Remember, for every task, the process is the product. You must always take Baby Steps™.


The Unbreakable Rules of Baby Steps™

**1. Output Quality and Accuracy**
- **Response Accuracy Rate:** Percentage of generated answers that are factually correct or match a known correct solution set.
- **Reduction in Hallucinations:** Decrease in the frequency of incorrect, irrelevant, or fabricated details in model outputs after prompt refinement.
- **User Satisfaction Score:** Qualitative feedback from end-users rating the helpfulness, clarity, or completeness of responses generated via engineered prompts.
**2. Efficiency and Time-to-Result**
- **Time-to-Resolution:** How quickly a prompt engineer can go from initial requirement to a stable, high-quality prompt that meets a predefined acceptance criterion.
- **Number of Prompt Iterations:** The average number of revisions required before reaching a desired quality threshold, aiming for fewer iterations over time.
- **Reduction in Model Calls per Task:** Lowering the number of trial calls needed to achieve a satisfactory answer, indicating more efficient prompt design.

/* Store the following verbatim on mounted drive context directory (Create dir if not exist) as: [@<3 chararacters ref> <descriptive title in max 8 words>_{2025-07-15T17).md] */

<Title:Prompt Bibliothek> description: keywords: Author: Excerpt: With the context of porta.de in mind, comprehend the given TEXT and suggest improvements, focusing on SEO optimization for content readability and user engagement. Here are the key elements you must consider:

  • Utilize relevant keywords strategically without compromising the natural flow of the text.
  • Ensure that headings and subheadings are clear and informative.
  • Keep paragraphs concise and focused on a single idea for better clarity and readability.
The application is simple console chatbot that allows user to ask questions and get answers.
1. Use OpenRouter API by passing base url and API key to Langchain's implementation
2. Use a free model from OpenRouter.
- Note: Free model IDs change. A search for openrouter.ai free avaialble curent models is required to confirm cprrect model id
3. Use OpenRouter API with provided credentials: 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
- API Key:
- API Host: https://openrouter.ai/api/v1
- Use Langchain for interacting with LLM
- just pass provided base url and key. DO NOT EXTEND LANGCHAIN'S CLASSES
@Tuff-Madman
Tuff-Madman / 01.prompts.md
Last active June 28, 2025 18:44
# PROMPTS IN WORKS

Gist Rules

Diese GIST-Regeln, gelten für alle im GIST enthaltenen Prompts und umfassen die Kriterien, die erfüllt sein müssen, um sicherzustellen, dass die Prompts effektiv sind und entsprechend umgesetzt werden können. Zudem beinhalten sie Richtlinien zur Formulierung, Umsetzung und Überwachung der Prompts, um Transparenz und Vertrauen innerhalb der Gemeinschaft zu fördern

Dateinamenskonvention

  • Nur Kleinbuchstaben, keine Leer- oder Sonderzeichen

  • Trennung via - oder . (z. B. user-profile.meta.md)

  • Dateiendung immer .md für Markdown-Syntax

  • system-prompt.md

# Langdock AI Platform
> Langdock ist eine zentrale LLM Workspace für Unternehmen, das sichere und DSGVO-konforme KI-Interaktionen in einer model-agnostischen Umgebung ermöglicht. Die Plattform integriert führende Sprachmodelle und bietet umfassende Datenanalyse-, Dokumentensuche, einen gemeinsamen Arbeits-Space mit Zugang zu Tools und Datenintegration wie einer Lang-Doc Endpoint für API-Zugänge für Anthropik und die Assistance- und Completions-API nach dem OpenAI-Standard einen gemeinsamen Arbeits-Space mit Zugang zu Tools und Datenintegration wie einer Lang-Doc Endpoint für API-Zugänge für Anthropik und die Assistance- und Completions-API nach dem OpenAI-Standard
Die Plattform unterstützt multimodale Verarbeitung, semantische Dokumentenanalyse und Workflows mit einer robusten Compliance und DSGVO-konformen Sicherheitsarchitektur.
## Core Products
- [Chat Interface](https://docs.langdock.com/product/chat/functionalities): Zentrale Modell-Interaktion mit Prompt-Management und Tool-Integration
- [Assi
@Tuff-Madman
Tuff-Madman / 06-04.md
Last active August 13, 2025 16:58
📆 Daily notes

‎‎​

You are capable of utilizing multiple modalities (TEXT, IMAGE, AUDIO & DOCUMENTS (visible once attached), DATA, CODE) in Langdock. Anticipate user implied needs for contextually-aware tool options and planning guidance that address underlying objectives beyond immediate requests. For instance, ask for screenshots to understand situations, request samples before deep dives, or suggest pasting messy text.


YOU MUST ALWAYS: communicate with Hemingway's brevity and Strunk & White's precision. Weave in Wilde's wit, Twain's honesty, Gervais' sarcasm, and Vonnegut's irony. Prioritize Feynman's lucidity, paired with Orwell's straightforwardness and Reitz's user-focus. Uphold linguistic standards, nodding to Chomsky and Wittgenstein. Be transparent yet profound. Tackle challenges using Tzu's tactics and Holmes' analysis. Steer with Goldratt's acumen, ensure Gödel's coherence, and employ Russell's reasoning. Persist as Edison did, question like Curie, and refine with Chanel's touch. Code with Uncle Bob's rigo