The Mission: We are not looking for a prompt engineer or a basic wrapper developer. We are building a high-performance, autonomous agent ecosystem. The goal is to deploy and orchestrate agents that can operate 24/7, utilize complex "Skills," and execute multi-step tasks across various platforms (Web, Airtable, WhatsApp, and Local File Systems). We need an architect who lives on the bleeding edge of the Agentic Era—specifically someone who can master OpenClaw and integrate it with general-purpose agents like Manus AI. Key Technical Requirements * OpenClaw Mastery: Deep experience with the OpenClaw/Molt framework, including Gateway configuration, Channel Adapters, and the Lane Queue system. * Skill Architecture: Proven ability to install, debug, and build custom Skills (YAML/Markdown-based). You must know how to give an agent "hands" via exec, browser-use, and API bridges. * Autonomous Orchestration: Expertise in setting up Heartbeats for proactive agent behavior and managing Multi-Agent handoffs. * General Agent Knowledge: Experience with Manus AI, OpenDevin, or similar autonomous agents for high-level task planning and execution. * Memory & State Management: Expertise in persistent memory strategies using JSONL transcripts, Semantic Snapshots, and Vector context. * Security & Isolation: Absolute focus on safety—managing Docker sandboxing and permission scoping for agents with shell/execution access. Core Responsibilities * Deploy and maintain a stable OpenClaw Gateway on a secure node environment. * Audit and install high-utility Skills from ClawHub or build custom internal tools. * Develop "Self-Healing" workflows where agents can detect errors in their own logic and retry with a new plan. * Integrate the agent ecosystem with our existing tech stack (Airtable, external APIs, and messaging channels). * Optimize token usage and API costs while maintaining high-frequency proactive checks (Heartbeats). How to Apply (The "Proof of Skill" Gatekeeper) Please start your application with the phrase: "AGENTIC ARCHITECT" In your proposal, please provide brief answers to the following: * Which OpenClaw Skills have you successfully implemented or customized? * How do you ensure an autonomous agent doesn't enter an "infinite loop" when a tool or skill fails? * Describe your preferred stack for running agents 24/7 (Docker? PM2? Cloud-based?). Thank you!