v1.0.0 - Initial Release: Omni-Channel Resale MCP Agent & Guardrail Hub

🚀 Initial Public Release of the SkillHub Resale Agent

This milestone marks the first official release of the resale-agent-skill-hub. Designed as a decentralized, local-first infrastructure, this project provides a production-ready Model Context Protocol (MCP) server and a modular skill framework to safely orchestrate C2C resale workflows.

🧬 Architectural Highlights

• Standardized MCP Server: Exposes 7 stateless tools via JSON-RPC over stdio. Client-agnostic design ensures seamless integration with modern LLM clients (Claude Desktop, Cursor, etc.) without vendor lock-in.
• Modular Skill Ecosystem: 8 independent, hot-pluggable Claude Code skills handling the entire lifecycle: vision-to-JSON parsing, multi-platform authentication, real-time pricing research, and automated scheduling.
• The Guardrail Hub (Pre-flight Gatekeeping): Implements a strict broker-fuse interceptor. This hard-stop middleware immediately revokes LLM publishing privileges if a generated price breaches the user-defined floor threshold.
• XAI Transparency Log: Features an append-only .jsonl audit trail. Every AI-driven decision (pricing, blocking, publishing) is recorded to ensure complete traceability and mitigate "black-box" liability.
• Zero-Friction Sandbox: Native BROKER_MOCK_MODE=true allows developers to safely evaluate the full 7-step pipeline and fuse interceptions without requiring real platform accounts or risking anti-scraping blocks.

🛠️ Quick Start

For full installation instructions across Claude Desktop and Claude Code CLI, please refer to the README.md.

This release establishes the baseline codebase for our ongoing open-source trust and guardrail framework initiatives.