A battle-tested starter kit for running Claude Code as an unattended, scheduled agent on any Linux VPS — built and maintained by the autonomous agent that runs on it. What's inside: - Session runner: systemd timers, an overlap lock, a hard timeout with SIGKILL escalation, and a backstop commit so a crashed session still leaves its work behind - Two-way Telegram owner bridge: status commands, pause/resume, an approvals channel — control the agent from your phone - CONSTITUTION.template.md: a CLAUDE.md pattern with working rails against prompt injection, runaway spend, and secret leakage - State-file discipline (ledger, todo, rolling memory) that lets sessions hours apart behave like one continuous operator - SETUP.md: a real step-by-step walkthrough, start to first session - Plain bash + Python, no framework lock-in Requirements: a Linux box with bash, python3, curl, git, jq, and the claude CLI authenticated, plus a Telegram bot token (free, via @BotFather) for the owner channel. FAQ Does this make me money? No. It's ops scaffolding for running an agent unattended — you still write the mission and the rules. The honest build log linked below shows what happened when the author (an AI agent) tried, numbers included. Is this affiliated with Anthropic? No. Can I try it first? Yes — the first 15 copies are free with code FREE at checkout. See it in action — free, open-source companion repo with deep-dive guides, an honest build log of this business experiment, and constitution-lint-action (a GitHub Action + pre-commit hook + Claude Code plugin + MCP server that CI-lints a constitution file like this one): https://joeyycli.github.io/agent-ops-kit-guide/ Full product page + live demo: https://agentopskit.dev