https://github.com/dejanu/ai-powered-coding-playbook
AI-powered development using resources from MCP to skills , agents and papers.
https://github.com/dejanu/ai-powered-coding-playbook
ai-tools anthropic antropic-claude claude-code tutorial
Last synced: 21 days ago
JSON representation
AI-powered development using resources from MCP to skills , agents and papers.
- Host: GitHub
- URL: https://github.com/dejanu/ai-powered-coding-playbook
- Owner: dejanu
- License: mit
- Created: 2026-01-11T13:51:12.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2026-05-10T12:40:43.000Z (about 1 month ago)
- Last Synced: 2026-05-10T13:44:58.968Z (about 1 month ago)
- Topics: ai-tools, anthropic, antropic-claude, claude-code, tutorial
- Language: Python
- Homepage:
- Size: 1.15 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# coding_AI
📍 Repo map:
* [anthropic_skilljar](https://github.com/dejanu/ai-powered-coding-playbook/blob/main/anthropic_skilljar/notes.md) resources
* [mcp demos](https://github.com/dejanu/ai-powered-coding-playbook/blob/main/mcp/notes.md) resources
📍 AI-enabled repo [taskapp](https://github.com/dejanu/taskapp)
## Papers
* [On the Opportunities and Risks of Foundation Models](https://arxiv.org/pdf/2108.07258)
* [Attention Is All You Need](https://arxiv.org/abs/1706.03762)
* [Improving Language Understanding by Generative Pre-Training](https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf)
* [Evaluating Large Language Models Trained on Code](https://arxiv.org/abs/2107.03374): Codex, a GPT language model fine-tuned on publicly available code from GitHub, and study its Python code-writing capabilities.
* [Sea Change in Software Development: Economic and Productivity Analysis of the AI-Powered Developer Lifecycle](https://arxiv.org/abs/2306.15033)
* [PREDICTING EXPERT EVALUATIONS IN SOFTWARE CODE
REVIEWS](https://arxiv.org/pdf/2409.15152): An algorithmic model that automates aspects of code review typically avoided due to their complexity
or subjectivity, such as assessing coding time, implementation time, and code complexity.
* [ReAct: Synergizing Reasoning](https://arxiv.org/abs/2210.03629)
---
## AI powered coding assistants:
Vanilla [Visual Studio Code] is built on top of Electron and benefits from all the networking stack capabilities of Chromium.
If your IDE is behind a proxy, check [Network Connections in Visual Studio Code](https://code.visualstudio.com/docs/setup/network).
[GitHub Copilot in Visual Studio Code](https://code.visualstudio.com/docs/copilot/reference/copilot-vscode-features): Powered by OpenAI Codex to work in real time, Codex is a GPT (more precisely a descendant of GPT-3) language model fine-tuned on publicly available code from GitHub.
[Cursor](https://cursor.com/) VSCode fork by Anysphere. Cursor [best practices](https://cursor.com/blog/agent-best-practices)
Install [Gemini Code in GitHub](https://developers.google.com/gemini-code-assist/docs/set-up-code-assist-github#install-gemini-code-assist) as code reviewer. After creation, Gemini Code Assist provides suggestions to your code review every time the pull request author or other human reviewers add comments with the `/gemini` tag on the pull request
## Terminology:
**Context** is everything:
Add files to chat aka `Mentions` via `#` (indexing and RAG search).
Add `chat participants` `@` (routes the question to the right agent or tool i.e. `@workspace how auth is implemented?` or `@terminal find the largest file in the src directory`)
* **Agents** autonomous session that follows the agentic loop of `undestand->act->validate`
* **Tools** are the specific functions or capabilities that agents can utilize to perform their tasks, such as asking questions, accessing APIs, reading files, searching the web, or executing commands.
* **Skills** playbooks/procedures for agents with step-by-step instructions on how to use tools to accomplish specific tasks, such as code review, debugging
* **Mentions** : type `@` in the chat input to attach specific context to your prompt. Start typing after `@` and Cursor shows matching suggestions.
```
# participants
@vscode how can I change key bindings
@terminal find the largest file in the src directory
```
## Agent examples in this repo
* [Reviewer](.github/agents/reviewer.agent.md): Review code for quality and adherence to best practices.
## Resources
* [Copilot-cheatsheet](https://docs.github.com/en/copilot/reference/cheat-sheet?tool=vscode)
* [Cursor keyboard shortcuts](https://cursor.com/docs/reference/keyboard-shortcuts)
* [Instruction Files](https://code.visualstudio.com/docs/copilot/customization/custom-instructions)
* [Chat Cookbook](https://docs.github.com/en/copilot/tutorials/copilot-chat-cookbook)
* [MCP](https://modelcontextprotocol.io/docs/getting-started/intro)
* [Prompt-cheatsheet](https://github.com/dejanu/coding_AI/blob/main/prompts/prompt-cheatsheet.md)
* [Google's AI tools](https://cloud.google.com/use-cases/free-ai-tools#overview)
---
# AI Tools
* Mathematical reasoning engine [Harmonic](https://harmonic.fun/)
* [Agent Skills](https://agentskills.io/home) (open standard)
* AI-powered research assistant [Undermind](https://www.undermind.ai/)
* LLM benchmarking [LiveBench](https://livebench.ai/#/?highunseenbias=true)