v1.2.0-skill: Packaged OpenClaw Skill

📦 Skill Package Release

This release provides the pre-packaged OpenClaw Skill for easy distribution and installation.

📥 Download

Skill Package (256KB):

• loc-mvr-v1.2.0.skill - Pre-packaged skill with all v1.2.0 features
• loc-mrv-v1.2.0.skill.sha256 - SHA256 checksum: 7a8873856fcdf9c66eff67a6fed25e5942cef2b73fb5b762300419f6ac4d47a9

🚀 Quick Install

# Download skill
wget https://github.com/Charpup/game-localization-mvr/releases/download/v1.2.0-skill/loc-mvr-v1.2.0.skill

# Verify checksum
sha256sum -c loc-mvr-v1.2.0.skill.sha256

# Extract and use
unzip loc-mvr-v1.2.0.skill
cd skill/
python scripts/normalize_guard.py examples/sample_input.csv ...

✨ Features Included

• Response Caching: 50%+ cost savings
• Intelligent Model Routing: 72% cost reduction
• Async/Concurrent Execution: 30-50% latency reduction
• Glossary AI System: 52% auto-approval rate
  • Smart Matcher
  • Auto-Corrector (with Russian declension support)
  • Learning System
  • Confidence Scoring (R² 0.88)
• Batch Optimization: 48% throughput increase

📊 Stats

• 27 production-ready scripts
• 24 test suites (1013+ tests)
• 91% code coverage
• Complete documentation

📚 Documentation

• English README
• 中文文档
• Full v1.2.0 Release Notes

v1.2.0-reorg: Clean Repository Structure

Repository Reorganization

This release features a complete cleanup of the repository structure for better maintainability.

What's New

Clean Structure:

• Removed numbered directories (01_active, 02_archive, etc.)
• Standard GitHub repository layout
• Organized src/, tests/, docs/, skill/, examples/
• Archived old logs to .archive/

No Functional Changes:

• All v1.2.0 features intact
• All tests passing
• Skill package unchanged

Downloads

Skill Package (recommended):

• loc-mvr-v1.2.0.skill - Pre-packaged OpenClaw skill
• SHA256: 7a8873856fcdf9c66eff67a6fed25e5942cef2b73fb5b762300419f6ac4d47a9

New Structure

game-localization-mvr/
├── README.md, README_zh.md
├── CHANGELOG.md, LICENSE
├── requirements.txt, Dockerfile
├── src/scripts/ - Core Python scripts
├── src/config/ - Configuration files
├── tests/unit/, tests/integration/
├── docs/ - Documentation
├── skill/ - Skill distribution
└── examples/ - Sample data

Quick Start

# Clone
git clone https://github.com/Charpup/game-localization-mvr.git

# Or download skill directly
wget https://github.com/Charpup/game-localization-mvr/releases/download/v1.2.0-reorg/loc-mvr-v1.2.0.skill
unzip loc-mvr-v1.2.0.skill

v1.2.0: Performance & Intelligence

🚀 Major Release

✨ New Features

🧠 Intelligent Model Router

• Complexity-based routing with 72% cost reduction
• Multi-factor analysis (text length, placeholders, glossary density)
• Automatic fallback and cost tracking

⚡ Async/Concurrent Execution

• 30-50% latency reduction
• 2-3x throughput increase (50-100 rows/sec)
• Streaming pipeline with backpressure handling

📚 Glossary AI System

• Smart glossary matcher: 52% auto-approval rate
• Auto-corrector with Russian declension support
• Learning system with weekly improvement
• Confidence scoring (R² 0.88 calibration)

🚀 Performance Improvements

• Response caching: 50%+ cost savings
• Batch optimization: 48% throughput increase
• 166x speedup in benchmarks
• 94.5% cost reduction at scale

📊 Stats

• 1013+ tests, 91% coverage
• 3200 lines documentation
• All targets exceeded!

📚 Documentation

• API Docs
• Configuration Guide
• Quick Start
• Performance Report

Loc-mvr Skill v1.1.0-patch2 - Production Ready

🔧 Hotfix Summary

Critical improvements to Docker usage guidance and troubleshooting.

This release supersedes v1.1.0 - Please use v1.1.0-patch2 for production.

---

🎯 What's New in v1.1.0-patch2

Critical Fixes (from v1.1.0)

• ✅ Docker Examples Added: Quick Start now includes container runtime examples (Rule 12 compliance)
• ✅ Trace Diagnostics: Issue 5 expanded with trace_config check commands
• ✅ Container Enforcement: New Issue 6 explains Docker requirement for LLM scripts
• ✅ Typo Fixes: Corrected "AQ"→"QA" and "Optmized"→"Optimized"

Improvements (from v1.1.0-patch1)

• 🧹 Clean Quick Start: Removed duplicate Docker examples (3→1)
• 🚀 Unified Workflow: Docker example now runs all 4 steps in single command
• 📝 Encoding Fix: Removed UTF-8 corruption in documentation

Enhancements

• 📦 README Badge: Direct Skill download link with verification guide
• 🐛 Issue Templates: Structured Bug Report and Feature Request forms

---

📦 Installation

Option 1: Download Pre-packaged Skill (Recommended)

# Download
wget https://github.com/Charpup/game-localization-mvr/releases/download/v1.1.0-patch2/loc-mvr-v1.1.0-patch2.skill

# Verify checksum
wget https://github.com/Charpup/game-localization-mvr/releases/download/v1.1.0-patch2/loc-mvr-v1.1.0-patch2.skill.sha256
sha256sum -c loc-mvr-v1.1.0-patch2.skill.sha256

# Extract and use
unzip loc-mvr-v1.1.0-patch2.skill
cd skill/

Option 2: Clone Full Repository

git clone https://github.com/Charpup/game-localization-mvr.git
cd game-localization-mvr
git checkout v1.1.0-patch2
pip install -r requirements.txt

---

🚀 Quick Start

Docker (All-in-One):

# Set credentials
export LLM_API_KEY="your_key"
export LLM_BASE_URL="https://api.example.com/v1"

# Run complete workflow in container
docker run --rm -v ${PWD}:/workspace -w /workspace \
  -e LLM_BASE_URL -e LLM_API_KEY \
  gate_v2 bash -c "
  python -u scripts/normalize_guard.py \
    examples/sample_input.csv output/normalized.csv \
    output/placeholder_map.json workflow/placeholder_schema.yaml &&
  python -u scripts/translate_llm.py \
    output/normalized.csv output/translated.csv \
    workflow/style_guide.md glossary/compiled.yaml &&
  python -u scripts/qa_hard.py \
    output/translated.csv output/qa_report.json \
    output/placeholder_map.json &&
  python -u scripts/rehydrate_export.py \
    output/translated.csv output/placeholder_map.json \
    output/final_export.csv
"

See SKILL.md for complete workflow (14 steps).

---

📊 Production Metrics (Verified at 30k rows)

Metric	Value	Target
Cost	$1.59/1k rows	<$2.00
Quality	<5 errors/30k	<1% error rate
Speed	~4h for 30k rows	-
Repair Rate	98.9% (Hard QA)	>95%

---

📚 Documentation

• SKILL.md: Complete workflow guide
• QA Rules: Hard QA logic & repair strategies
• Glossary Spec: Term management lifecycle
• Metrics Guide: Cost optimization formulas

---

🐛 Reporting Issues

Use our Bug Report Template for structured issue reporting.

---

🙏 Credits

Built with Claude for game localization automation.

v0.5.0 - Feature Pack E+F

What's New

Feature Pack E: Strict Prompts

• translate_llm.py: Strict pure text output
• soft_qa_llm.py, glossary_*, style_guide_score.py: Strict JSON schema
• Punctuation control: 【】 → «»

Feature Pack F: Concurrency Fix (P0)

• translate_llm.py: ThreadPoolExecutor + as_completed
• CLI arg: --max_inflight (default=3)
• Metrics: queue_wait_ms, inflight_limit in LLM trace

New Tools

• lib_text.py, llm_prompt_inventory.py, dev_prompt_improver.py
• diff_translation.py, normalize_ingest.py