Agent Skills for Context Engineering - teaching agents to "think contextually"

This repository shows how to enhance LLM agents so that they better understand the task, the dialogue history, and the conditions, rather than just generating responses.

Useful for:

• skills in managing long contexts
• neat structuring of data and instructions
• templates for searching, filtering, and making decisions
• examples of real scenarios (chats, memory tasks, integrations)

Often, agents lose important details, confuse steps, and "forget" the goal. This library teaches them to keep the context under control and act more consistently.

Github

••••••••••••••••••••••••••••••••••••••••••••••••••••

✅ 10 ChatGPT Prompts To Save You Hours of Work ⏳🤖

1️⃣ Content Writing
"Act as a blog writer. Write a 600-word blog post on ‘5 AI tools boosting productivity in 2025’ with SEO keywords."

2️⃣ YouTube Scripts
"Act as a video scriptwriter. Write a 2-minute YouTube Shorts script explaining how ChatGPT works."

3️⃣ Online Course Creation
"Act as an edtech consultant. Help me outline a 5-module beginner course on Python for absolute newbies."

4️⃣ Portfolio Building
"Act as a career mentor. Suggest 5 project ideas I can add to my GitHub to showcase my web development skills."

5️⃣ Newsletter Writing
"Act as a newsletter editor. Write a weekly tech newsletter for young professionals learning AI."

6️⃣ Presentation Design
"Act as a slide designer. Create a bullet-point outline for a pitch deck introducing a startup AI tool."

7️⃣ Freelance Proposals
"Act as a freelancer. Write a proposal to apply for an Upwork gig: ‘Build a landing page in HTML/CSS in 3 days.’"

8️⃣ Idea Validation
"Act as a startup mentor. Evaluate the business potential of an AI-based resume review tool."

9️⃣ Client Communication
"Act as a client success manager. Write a professional response to a client unhappy with project delays."

🔟 Learning Tracker
"Act as a mentor. Design a Notion template to track progress for someone learning data science in 3 months."

💬 Tap ❤️ for more!

🤖 12 AI Terms to Learn 👇🏻

🎯If you want to survive in AI era, you must complete these 5 Free AI Courses by google before the 2026 ends 👇

1/ Introduction to Generative AI:
https://www.skills.google/course_templates/536

2/ Introduction to LLM:
https://www.skills.google/course_templates/539

3/ Introduction to Responsible AI:
https://www.skills.google/course_templates/554

4/ GenAI Bootcamp:
https://cloudonair.withgoogle.com/gen-ai-bootcamp

5/ Google AI Essentials:
https://www.skills.google/paths/2336

The 10 Most Valuable AI Learning Repositories on GitHub 👇

I pulled the top 10 repos where Jupyter is the main language
Filtered for the best educational resources

Here's what's worth your time :

1. microsoft/generative-ai-for-beginners ⭐ 105,577 21
lessons covering the full GenAI stack From prompting basics to production deployment Built by Microsoft's AI education team
🔗 https://lnkd.in/diW9Cca6

2. rasbt/LLMs-from-scratch ⭐ 83,714
Build GPT-like models from zero No hand-waving, pure implementation Companion to Sebastian Raschka's book
🔗 https://lnkd.in/d3cq5diH

3. microsoft/ai-agents-for-beginners ⭐ 49,333
Complete course on agentic systems Covers planning, tools, memory, multi-agent Released 3 months ago, already essential
🔗 https://lnkd.in/e-a2gqSv

4. microsoft/ML-For-Beginners ⭐ 83,279
12 weeks of classical ML fundamentals 26 lessons, 52 quizzes, full curriculum Still relevant despite the LLM hype
🔗 https://lnkd.in/e7S8yDbS

5. openai/openai-cookbook ⭐ 71,106
Official OpenAI examples and guides Real production patterns, not toys Updated constantly with new features
🔗 https://lnkd.in/dtMbuMGk

6. jackfrued/Python-100-Days ⭐ 177,958
Most-starred educational repo on GitHub 100 days from Python beginner to advanced Covers web dev, data science, automation
🔗 https://lnkd.in/duWVtn4i

7. pathwaycom/llm-app ⭐ 54,583
Production RAG templates you can deploy Real-time data pipelines, not static demos Enterprise search with live updates
🔗 https://lnkd.in/daUFK9Nd

8. jakevdp/PythonDataScienceHandbook ⭐ 46,574
Entire data science handbook as Jupyter notebooks NumPy, Pandas, Matplotlib, Scikit-Learn Free alternative to $60 textbook
🔗 https://lnkd.in/db8HP7vT

9. CompVis/stable-diffusion ⭐ 72,246
Original Stable Diffusion implementation Understand how text-to-image actually works Foundation for SDXL, Midjourney competitors
🔗 https://lnkd.in/dEya2Rb5

10. facebookresearch/segment-anything ⭐ 53,250
Meta's SAM model for computer vision Promptable segmentation in images and videos Powers modern AI video editing tools
🔗 https://lnkd.in/dKvjk6Yb

✅ Amazing AI Facts You Should Know 🤖🧠

1. AI can now write code, generate art, and compose music using generative models like Chat and DALL·E.
2. The term "Artificial Intelligence" was coined in 1956 at the Dartmouth Conference.
3. AI powers most online recommendations (Netflix, YouTube, Amazon, etc.).
4. Self-driving cars use AI to detect lanes, traffic signs, and obstacles in real-time.
5. Machine Learning is a subset of AI, where systems learn from data without being explicitly programmed.
6. AI is used in healthcare to detect diseases, predict risks, and assist surgeries.
7. Deep Learning uses neural networks inspired by the human brain.
8. AI voice assistants like Siri, Alexa, and Google Assistant use NLP (Natural Language Processing).
9. Generative AI can create fake but realistic images, voices, and even videos (deepfakes).
10. AI can beat world champions in chess (Deep Blue) and Go (AlphaGo).
11. AI ethics is a growing field, focusing on fairness, transparency, and bias reduction.
12. Chat reached 1 million users in just 5 days, faster than any app in history.
13. Most smartphones already use AI for camera optimization, battery management, and face unlock.
14. AI is helping scientists discover new drugs faster using protein structure prediction (AlphaFold).
15. AI is predicted to impact over 80% of jobs, reshaping industries in the next decade.

💬 Tap ❤️ for more!

Now, let’s understand another AI Project:

🚀 Project 6: AI Resume Screening System

🎯 Problem Statement

Automatically screen resumes and rank candidates based on job description

Example:
• Input: Resume + Job Description
• Output: Match Score + Ranking

🧠 What You’re Building

👉 A system that:
• Reads resumes (PDF/Text)
• Extracts skills
• Compares with job description
• Gives matching score

📊 Dataset

Use:
• Resume dataset (Kaggle)
• Or create your own sample resumes

Format:
Resume Text | Skills | Category

⚙️ Step 1: Extract Text from Resume

👉 If PDF:

import PyPDF2

def extract_text(file):
    reader = PyPDF2.PdfReader(file)
    text = ""
    for page in reader.pages:
        text += page.extract_text()
    return text

🧹 Step 2: Text Preprocessing
• Lowercase
• Remove symbols
• Tokenization

🔢 Step 3: Convert Text → Features

👉 Use TF-IDF

from sklearn.feature_extraction.text import TfidfVectorizer

vectorizer = TfidfVectorizer()

🤖 Step 4: Similarity Calculation

👉 Compare resume vs job description

from sklearn.metrics.pairwise import cosine_similarity

score = cosine_similarity(resume_vec, jd_vec)

📊 Step 5: Ranking System

👉 Rank candidates based on score

🌐 Step 6: Build UI (Streamlit)

Features:
• Upload resume
• Enter job description
• Show match score

📁 Project Structure
resume-screening/
│
├── app.py
├── model.py
├── utils.py
├── requirements.txt
├── README.md

📝 Resume Description

AI Resume Screening System
• Built NLP-based system to match resumes with job descriptions
• Used TF-IDF and cosine similarity for ranking candidates
• Extracted text from PDFs and processed using NLP techniques
• Developed interactive app using Streamlit

🎯 Skills You Show

✔ NLP
✔ Feature extraction
✔ Similarity algorithms
✔ Real-world AI system
✔ Deployment

🔥 Make It 10/10 Project

Add:
✔ Multiple resume upload
✔ Skill extraction (NER)
✔ Top candidate ranking
✔ Dashboard

⚠️ Common Mistakes

❌ Only comparing text directly
❌ No preprocessing
❌ No ranking logic
❌ No UI

👉 This project shows:
• Real business problem solving
• Automation mindset
• Practical NLP

🚀 Double Tap ❤️ For More

Now, let’s understand another AI Project:

🚀 Project 7: End-to-End AI Assistant (Multi-Feature App 🔥)

This single project can replace 3–4 basic ones if done properly.

🎯 Problem Statement

Build an AI Assistant App that can:
- Answer questions (Chatbot)
- Analyze text (Sentiment)
- Summarize content
- (Optional) Answer questions from PDF

👉 One app → multiple AI features

🧠 What You’re Building

A multi-functional AI system combining:

✔ NLP
✔ Generative AI
✔ ML
✔ Deployment

⚙️ Tech Stack
- Python
- OpenAI / Hugging Face
- Scikit-learn
- Streamlit

🔹 Core Features (Must Have)

💬 1. Chatbot
- Ask anything → get response

😊 2. Sentiment Analyzer
- Input text → Positive/Negative

📝 3. Text Summarizer
- Long text → short summary

📄 4. PDF Q&A (Advanced 🔥)
- Upload PDF
- Ask questions

🔹 Step-by-Step Approach

1️⃣ Build Chatbot

Use LLM API:
response = client.chat.completions.create(...)

2️⃣ Add Sentiment Model

Reuse your sentiment project

3️⃣ Add Summarization

Use LLM:
"Summarize this text..."

4️⃣ Add PDF Feature (Optional)
- Extract text
- Use LLM to answer

5️⃣ Build UI (Streamlit)

👉 Tabs for each feature:
- Chat
- Sentiment
- Summary
- PDF

📁 Project Structure
ai-assistant/
│
├── app.py
├── chatbot.py
├── sentiment.py
├── summarizer.py
├── requirements.txt
├── README.md

🌐 Deployment

👉 Must deploy this

Use:
- Streamlit Cloud
- Hugging Face Spaces

📝 Resume Description

AI Assistant Application
- Built multi-feature AI app including chatbot, sentiment analysis, and text summarization
- Integrated LLM APIs for dynamic content generation
- Developed interactive UI using Streamlit
- Designed modular system combining multiple AI functionalities

🎯 Skills You Show

✔ Generative AI
✔ NLP
✔ System design
✔ API integration
✔ Deployment

🔥 Why This Project is Powerful

👉 Shows:
- You can combine multiple AI concepts
- You can build real-world applications
- You understand modern AI

⚠️ Common Mistakes

❌ Only chatbot
❌ No structure
❌ No UI
❌ No deployment

🧠 Pro Tip

👉 Keep it:
- Simple
- Clean
- Working

👉 Don’t overcomplicate

🏁 Double Tap ❤️ For More