polis-protocol

Set up a self-optimizing multi-vendor AI agent team using the Polis Protocol — a markdown-based "small city" where each agent is a citizen with a content-hashed capability card, tasks are structured contracts routed to whoever has the strongest track record by a learning bandit, settled contracts produce lessons that compound into team memory, and citizens can ratify amendments to the protocol itself. Use whenever several AI agents (Anthropic, OpenAI, Google, or any vendor) collaborate on one project and "who should do this" is a real question; when you want work routed to whichever model is best at it; when you want a team that measurably gets better over time; or to set up a new polis, register or join as a citizen, open, claim, or settle a contract, file a lesson, propose or vote on an amendment, run a chavruta review before a high-stakes irreversible action, troubleshoot a stalled contract or router pathology, override a routing recommendation, or migrate from a simpler shared-vault setup (e.g. agent-vaul

Set up a self-optimizing multi-vendor AI agent team using the Polis Protocol — a markdown-based "small city" where each agent is a citizen with a content-hashed capability card, tasks are structured contracts routed to whoever has the strongest track record by a learning bandit, settled contracts produce lessons that compound into team memory, and citizens can ratify amendments to the protocol itself. Use whenever several AI agents (Anthropic, OpenAI, Google, or any vendor) collaborate on one project and "who should do this" is a real question; when you want work routed to whichever model is best at it; when you want a team that measurably gets better over time; or to set up a new polis, register or join as a citizen, open, claim, or settle a contract, file a lesson, propose or vote on an amendment, run a chavruta review before a high-stakes irreversible action, troubleshoot a stalled contract or router pathology, override a routing recommendation, or migrate from a simpler shared-vault setup (e.g. agent-vaul

Set up a self-optimizing multi-vendor AI agent team using the Polis Protocol — a markdown-based "small city" where each agent is a citizen with a content-hashed capability card, tasks are structured contracts routed to whoever has the strongest track record by a learning bandit, settled contracts produce lessons that compound into team memory, and citizens can ratify amendments to the protocol itself. Use whenever several AI agents (Anthropic, OpenAI, Google, or any vendor) collaborate on one project and "who should do this" is a real question; when you want work routed to whichever model is best at it; when you want a team that measurably gets better over time; or to set up a new polis, register or join as a citizen, open, claim, or settle a contract, file a lesson, propose or vote on an amendment, run a chavruta review before a high-stakes irreversible action, troubleshoot a stalled contract or router pathology, override a routing recommendation, or migrate from a simpler shared-vault setup (e.g. agent-vaul