What's New in v1.4.0
🤖 AI-Powered Code Review Pipeline
- 4 GitHub Actions workflows for fully automated code review and fix cycle
ai-review.yml— CRITICAL/WARNING severity split for targeted reviewauto-merge.yml— @copilot auto-fixes WARNING-level issuesdecompose.yml— Context Sufficiency Check before task decompositionDESIGN_CONSTRAINTS.mdinjection into review context
🧩 7 New Skills
| Skill | Purpose |
|---|---|
| coding-agent-router | Route tasks to the best coding agent |
| context-and-impact | Context engineering + impact analysis |
| copilot-coding-agent | GitHub Copilot integration |
| copilot-full-automation | End-to-end Copilot pipeline |
| cursor-agent | Cursor AI integration |
| devin-agent | Devin AI integration |
| manus-agent | Manus AI integration |
Plus: X auto-posting scheduler, note.com publisher, ScrapeCreators API wrapper (27 platforms), last30days research integration.
🔧 Quality Improvements
- Flagged threshold raised from 0.7 → 0.8 (stricter quality monitoring)
- Dead links fixed (
agentskills.bath.me→agentskills.io) - Incorrect Discord link removed
- CI authentication switched to Claude Code OAuth
📊 Stats
- ⭐ 150 stars and growing
- 42 agents in daily production use
- 57% fewer skill failures with self-improving loop enabled
Full Changelog: v1.3.0...v1.4.0