v0.1.6

Added

• Added Weaviate-backed semantic vector storage for RAG while keeping BM25 local.
• Added RAG cache bundle export/import workflow for distributing prebuilt knowledge-base indexes.
• Added a retrieval-only LLM query rewrite step that rewrites user input into frontend-oriented search queries before RAG.

Changed

• Clarified all RAG-facing terminology so remote RAG evidence is referred to as "knowledge base" instead of the current workspace repository.
• Updated English and Chinese documentation with Weaviate, query rewrite, and cache bundle usage examples.

Fixed

• Prevented the planner from treating remote RAG hits as local workspace files during query tasks.
• Improved semantic index resilience for Weaviate-backed retrieval and related RAG execution flow.

v0.1.5

Full Changelog: v0.1.4...v0.1.5

v0.1.4

What's Changed

• feat[ptr]: 将规划者的可复用能力全部技能化，planner 通过技能注册表选择执行skill，不再直接硬编码 by @ceilf6 in #8

Full Changelog: v0.1.3...v0.1.4

v0.1.3

What's Changed

• release: 0.1.3 by @ceilf6 in #7

New Contributors

• @ceilf6 made their first contribution in #7

Full Changelog: v0.1.0...v0.1.3

v0.1.2 - npm Package Fix

🐛 Bug Fixes

This release fixes the npm installation issues from v0.1.0 and v0.1.1.

What's Fixed

• ✅ Bundled all dependencies into a single executable using esbuild
• ✅ Reduced package size from 709KB to 6MB (bundled)
• ✅ Fixed module resolution errors when installing from npm
• ✅ Simplified installation - now only requires playwright as external dependency

📦 Installation

# Install globally
npm install -g frontagent

# Or using pnpm
pnpm add -g frontagent

# Or using npx (no installation needed)
npx frontagent --help

🚀 Quick Start

# 1. Configure your LLM provider
export PROVIDER="openai"
export API_KEY="sk-..."
export MODEL="gpt-4"

# 2. Navigate to your project and initialize
cd your-project
frontagent init

# 3. Run tasks
frontagent run "Create a user login page"

📋 Package Info

• npm: https://www.npmjs.com/package/frontagent
• Size: 6.0 MB (compressed), 35.7 MB (unpacked)
• Files: 5 (bundled)

📝 Changes Since v0.1.0

• Added esbuild bundling for CLI
• Removed workspace dependency issues
• Updated installation documentation
• Added npm badges to README

---

Full Changelog: v0.1.0...v0.1.2

v0.1.0 - Initial Release

🎉 Initial Release

FrontAgent is an enterprise AI agent platform for frontend engineering with planning/execution architecture, SDD constraints, and MCP-controlled code generation.

✨ Features

• Planning/Execution Architecture: Dual-stage architecture for reliable code generation
• SDD-Based Constraints: Software Design Document driven development workflow
• MCP Protocol Integration: Model Context Protocol for controlled actions
• Project Automation: Full project-level automation support
• Browser Validation: Built-in browser port validation
• File Structure Scanning: Pre-planning project structure analysis

📦 Installation

npm install -g frontagent@0.1.2
# or
pnpm add -g frontagent@0.1.2

Note: Please use v0.1.2 instead, which includes bug fixes for npm installation.

📋 Requirements

• Node.js >= 20.0.0
• pnpm >= 9.0.0

---

Full Changelog: https://github.com/ceilf6/FrontAgent/commits/v0.1.0