New to both projects? Start with the copyable ChatGPT learning prompt. It teaches KGO and the read-only Overseer Hub example step by step, asks questions, practises with your own prompts, and generates a portable Learning Passport for future chats. No installation is required to begin.
One maintained version of the Kimi Genesis Orchestrator (KGO) skill, supplied for learning and installing into your own AI agent.
Start with the README in kimi-genesis-orch/. It contains a copyable teaching/setup prompt, installation instructions, an example exercise, and a verification checklist.
- Core skill — read this in full first. It only activates by explicit request.
- Strategy catalog — read the complete entries relevant to your task.
- Failure modes — common mistakes and how to avoid them.
- Worker-prompt slots — use only when you want to write instructions for another agent.
To get the files, use GitHub Code → Download ZIP, or run git clone https://github.com/pykmintin/Repo-share.git. Keep the complete kimi-genesis-orch/ directory together.
This release preserves the canonical SKILL.md, three reference files, and agents/openai.yaml byte-for-byte. The README is a portable onboarding adaptation; maintainer-specific retrieval configuration is omitted. Historical v3/v4 editions were retired from this public repository and archived offline.
No API keys or special orchestration backend are needed to read and understand KGO. Access to files, tool calls, and other agents depends on your chosen AI client.
Open the NotebookLM starter for a copyable agent prompt, local installation scripts, normal Google browser login, server-verified authentication checks and an optional cookie keepalive. One account is the default; independent named profiles support multiple accounts if you need them. This is a self-contained public example and needs none of the maintainer's private accounts, Overseer Hub or KeyHub services. It installs the unofficial notebooklm-py CLI with the upstream browser integration instead of distributing Google credentials or copied upstream code.
Read the Overseer Hub starter guide to learn and deploy a separate, read-only MCP backend using your own Auth0 tenant, HTTPS hostname and ChatGPT custom MCP connection. The architecture explanation maps it to the real Overseer design. It includes source, environment template and executable tests; it does not include the private production Bridge, browser sessions, host tools or credentials.