{
  "markdown": "# MO§ES™ Governance — Cowork & Chat Edition\n\nConstitutional AI governance for Claude.ai Chat and Cowork. Prompt-native — no installation, no Python, no shell dependencies.\n\n**© 2026 Ello Cello LLC** | Patent pending: Serial No. 63/877,177 | [mos2es.io](https://mos2es.io)\n\n---\n\n## What It Does\n\nBrings the MO§ES™ governance framework to Claude.ai without any runtime dependencies. Set governance modes, operational postures, and role hierarchies — Claude enforces them behaviorally for the rest of the conversation.\n\nWorks in **Claude.ai Chat**, **Cowork projects**, and anywhere the Claude plugin system is available.\n\n---\n\n## Commands\n\n| Command | What it does |\n|---------|-------------|\n| `/govern [mode]` | Set governance mode — applies behavioral constraints immediately |\n| `/posture [posture]` | Set operational posture — controls transaction scope |\n| `/role [role]` | Set agent role — controls authority in multi-agent workflows |\n| `/vault` | Load a governance document into active context |\n| `/audit` | View session governance audit trail |\n| `/status` | Show full current governance state |\n\n---\n\n## Governance Modes\n\n| Mode | Purpose |\n|------|---------|\n| `high-security` | Verify all claims, require confirmation before destructive actions |\n| `high-integrity` | Accuracy above speed, cite sources, flag uncertainty |\n| `creative` | Explore freely, log when shifting to speculation |\n| `research` | Deep investigation, document methodology and provenance |\n| `self-growth` | Reflective learning, track improvements |\n| `problem-solving` | Decompose before solving, verify against requirements |\n| `idk` | Guided discovery, clarifying questions, explicit tradeoffs |\n| `unrestricted` | No constraints, session audit still active |\n\n---\n\n## Operational Postures\n\n| Posture | Scope |\n|---------|-------|\n| `scout` | Read-only — gather, analyze, report. No state changes. |\n| `defense` | Protect existing positions. No new commitments. |\n| `offense` | Full scope — create, modify, transact, expand. |\n\n---\n\n## How Enforcement Works\n\nThis plugin uses **prompt-native enforcement** — Claude applies governance constraints by following the skill and command instructions. There are no shell hooks or runtime processes.\n\nEnforcement model:\n- Mode constraints are injected into Claude's active context when you run `/govern`\n- Claude self-enforces them for the remainder of the conversation\n- `/audit` reconstructs the governance event log from conversation history\n- Session state resets when the conversation ends\n\n**For cryptographic audit chains, hook-level enforcement, and persistent state across sessions**, use the Claude Code plugin: [moses-governance](https://github.com/SunrisesIllNeverSee/moses-governance)\n\n---\n\n## Two Plugins, One Framework\n\n| | Claude Code Plugin | Cowork/Chat Plugin |\n|---|---|---|\n| **Enforcement** | Hook-level (automatic) | Prompt-native (behavioral) |\n| **Audit** | SHA-256 cryptographic chain | Conversation-native |\n| **State** | Persistent JSON on disk | Session-scoped |\n| **Runtime** | Python + shell | None |\n| **Install** | Claude Code marketplace | Claude.ai plugins |\n\nSame constitutional framework. Different runtime targets.\n\n---\n\n## Framework\n\nBuilt on MO§ES™ — a constitutional governance framework for AI agents. Governance rules are encoded as machine-readable constraints, not just guidelines. The framework includes mode definitions, posture controls, role hierarchies, vault document injection, and amendment procedures.\n\nPreprint published. Patent pending: PPA4, Serial No. 63/877,177.\n\n---\n\n## License\n\nSource-available for evaluation. Commercial license: [mos2es.io](https://mos2es.io)\n\n© 2026 Ello Cello LLC\n",
  "bytes": 3676,
  "sha": "107a51bb1680b6d227bf29313bfe3a3419141e6988c106ec467a9d72cf97dd4c",
  "repo_slug": "sunrisesillneversee/moses-governance-cowork",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_sunrisesillneversee_moses_governance_cow_e054686e/readme"
}