What's New in v3.0.4
Universal Installation Method
- NEW: Direct Download Option for developers without Node.js
- Works with any programming language: Python, Java, Go, Rust, PHP, Ruby, C++, etc.
- Works with any IDE: VS Code, IntelliJ, PyCharm, Vim, etc.
- Simple curl command for quick setup
- Manual download instructions from GitHub releases
Installation Options
Option 1: NPM (Node.js projects)
npx ctxforge init
Option 2: Direct Download (Universal)
curl -L https://github.com/vencolini/ctxforge/archive/refs/heads/main.zip -o ctxforge.zip
unzip ctxforge.zip
cp -r ctxforge-main/lib/framework docs/context
rm -rf ctxforge.zip ctxforge-main
Documentation Improvements (from v3.0.3)
- Added ctxforge logo to README for better visual identity
- Updated subtitle to "Lightweight context discovery and engineering framework"
- Added comprehensive "What Problem Does This Solve?" section
- Added "Inspiration" section with link to https://www.youtube.com/watch?v=LCEmiRjPEtQ&t=1102s
v3.0 Major Features
Protocol Auto-Loading System
- 60-70% Token Reduction: 8-10K tokens per session vs 25K in v2.1
- Intent Detection: Automatically loads the right protocol based on your request
- 15 Specialized Protocols: Feature Development, Bug Fixing, Refactoring, Code Review, Testing, Investigation,
Documentation, Performance Optimization, Security Audit, Architecture Design, Dependency Management, Deployment, Pair
Programming, Learning, Database Migration
- Unlimited Scalability: Add new protocols without increasing overhead
- Multi-Protocol Sessions: Switch between different workflows in the same session
How It Works
1. LLM reads CORE.md (5K tokens)
2. You describe what you want to work on
3. Intent detection table identifies the task type
4. LLM loads only the relevant protocol (3-5K tokens)
5. Protocol guides discovery → implementation → documentation
Pain Points Addressed
- Context Overload: Loading massive instruction sets (25K+ tokens) that leave little room for your actual code
- Premature Implementation: LLMs jumping to code before understanding requirements
- Missing Edge Cases: Features built without discovering critical constraints or error scenarios
- Repeated Mistakes: No systematic way to capture and apply project-specific learnings
- Inconsistent Quality: No automatic enforcement of performance, security, and accessibility standards
Links
- NPM: https://www.npmjs.com/package/ctxforge
- Inspiration: https://www.youtube.com/watch?v=LCEmiRjPEtQ&t=1102s
Full Changelog: https://github.com/vencolini/ctxforge/compare/v3.0.3...v3.0.4
