MO§ES™ Posture Controls
Posture is the throttle. Mode is the rulebook. Both always apply.
Load active posture from ~/.openclaw/governance/state.json before any action that touches state, executes a transaction, or makes an external call.
Dependency: This skill reads ~/.openclaw/governance/state.json (declared in stateDirs) and invokes init_state.py from the moses-governance skill bundle (declared in requires). Install moses-governance alongside this skill. No secrets or credentials are required.
SCOUT — Read-Only
Behavior: Information gathering only.
Transaction policy: NO transactions. NO state changes. NO file writes.
What you can do: Read, analyze, report, recommend.
What you cannot do: Execute trades. Transfer funds. Write files. Call APIs that mutate state.
When operator tries to execute in SCOUT: Inform them — "SCOUT posture active. No execution permitted. Switch to DEFENSE or OFFENSE to proceed."
Use when: Initial assessment, due diligence, market analysis before committing.
DEFENSE — Protect
Behavior: Protect existing positions.
Transaction policy: Outbound transfers require explicit operator confirmation. Double confirmation for transfers exceeding 10% of position.
What you can do: Read, analyze, execute inbound/neutral operations, small confirmed outbound.
What you must do: Flag any action that reduces holdings. Require operator to type "CONFIRM [action]" before executing outbound transfers.
Use when: Volatile conditions, risk management, custody ops, protecting assets.
OFFENSE — Execute
Behavior: Execute on opportunities.
Transaction policy: Permitted within governance mode constraints. All executions logged with rationale.
What you can do: Full execution within mode constraints.
What you must do: Log every execution decision with rationale. Track performance.
Use when: Active trading, deployment, operator has assessed risk and is ready to act.
Posture + Mode Interaction
| Mode + Posture | Result |
|---|
| High Security + SCOUT | Maximum caution. Read-only. Every data point verified before reporting. |
| High Security + DEFENSE |
Outbound blocked without verification + explicit confirmation. |
| High Security + OFFENSE | Executes with full verification chain + confirmation required. |
| High Integrity + SCOUT | Deep verified research. No execution. |
| Creative + SCOUT | Explore ideas freely. No execution. |
| Creative + OFFENSE | Experimental execution — audited. |
| Research + SCOUT | Gather everything. Act on nothing. |
| Unrestricted + OFFENSE | Full autonomy. Logged. Operator accepts all risk. |
MO§ES™ 姿态控制
姿态是油门。模式是规则手册。两者始终适用。
在执行任何涉及状态、执行交易或进行外部调用的操作之前，从 ~/.openclaw/governance/state.json 加载当前姿态。
依赖关系： 此技能读取 ~/.openclaw/governance/state.json（在 stateDirs 中声明），并从 moses-governance 技能包（在 requires 中声明）调用 init_state.py。请将此技能与 moses-governance 一同安装。无需密钥或凭证。
侦察模式 — 只读
行为： 仅收集信息。
交易策略： 禁止交易。禁止状态变更。禁止文件写入。
可执行操作： 读取、分析、报告、建议。
禁止操作： 执行交易。转移资金。写入文件。调用会改变状态的API。
当操作员试图在侦察模式下执行操作时： 告知他们——当前为侦察姿态。禁止执行。请切换至防御或进攻模式以继续。
适用场景： 初始评估、尽职调查、承诺前的市场分析。
防御模式 — 保护
行为： 保护现有头寸。
交易策略： 对外转账需操作员明确确认。超过头寸10%的转账需双重确认。
可执行操作： 读取、分析、执行入站/中性操作、小额已确认的对外操作。
必须执行的操作： 标记任何减少持仓的操作。要求操作员在执行对外转账前输入确认 [操作]。
适用场景： 波动行情、风险管理、托管操作、资产保护。
进攻模式 — 执行
行为： 执行机会操作。
交易策略： 在治理模式约束范围内允许执行。所有执行均需记录理由。
可执行操作： 在模式约束范围内完全执行。
必须执行的操作： 记录每次执行决策及其理由。跟踪绩效。
适用场景： 主动交易、部署、操作员已评估风险并准备行动。
姿态与模式交互
| 模式 + 姿态 | 结果 |
|---|
| 高安全 + 侦察模式 | 最大谨慎。只读。每个数据点在报告前均需验证。 |
| 高安全 + 防御模式 |
未经验证+明确确认，禁止对外操作。 |
| 高安全 + 进攻模式 | 执行需经过完整验证链+确认。 |
| 高完整性 + 侦察模式 | 深度验证研究。禁止执行。 |
| 创意 + 侦察模式 | 自由探索想法。禁止执行。 |
| 创意 + 进攻模式 | 实验性执行——需审计。 |
| 研究 + 侦察模式 | 收集一切。不采取任何行动。 |
| 无限制 + 进攻模式 | 完全自主。记录日志。操作员承担所有风险。 |
