v1.1.0 — CLI, MCP Server & Multi-Provider Expansion

What's New

Architecture Refactor

• core.py — Extracted data layer (schema validation, MongoDB CRUD) with no UI dependencies
• app.py — Refactored to import from core; UI-only

Command-Line Interface (cli.py)

• ingest — Validate and store fragments (human or machine) from the terminal
• retrieve — Fetch random fragments with filters (origin, model, language, veracity)
• stats — Dataset statistics
• models — List all configured model identifiers
• Strict/lenient model validation against prompt_engine.json

MCP Server (mcp_server.py)

• Model Context Protocol integration for AI agent workflows
• Tools: ingest_fragment, retrieve_fragments
• Resources: roguegpt://config/models, roguegpt://config/languages, roguegpt://stats

Model Configuration

• Expanded from 7 to 37 model configurations across 10 providers
• New providers: Anthropic, DeepSeek, Zhipu, Moonshot, Qwen, MiniMax
• New models include GPT-4.1, o1/o3 series, Claude Opus 4.6, Gemini 3 Pro, LLaMA-3.3 70B, DeepSeek R1/V3, and more

Documentation

• Complete README rewrite with architecture diagram, fragment schema, CLI/MCP usage examples
• All three citation entries (WWW '26 Companion ×2, foundational survey)

Dataset

• 2,308 fragments (51 human, 2,257 machine) across 4 languages and 3 formats
• Zenodo archive: 10.5281/zenodo.18703138