An open protocol and engine for human-AI collaboration, governed by the CORTEX Six Laws. 面向人机协作的开源协议与引擎,由 CORTEX 六定律治理。
| Law | English / 中文 |
|---|---|
| C1 | Intent as Boundary — human intent is the sole boundary of AI action; unauthorized action is out of bounds; ask rather than assume when intent is unclear. 意图为界 — 人类意图是 AI 行动的唯一边界;未授权即越界;意图不清时询问而非假设。 |
| C2 | Trust as Measure — degree of autonomy is set by verifiable trust, accumulated through track record, not granted as a binary switch. 信任为尺 — 自主程度由可验证的信任决定,通过历史表现积累,不是二元开关。 |
| C3 | Safety as Foundation — reversible actions can proceed freely; irreversible actions require caution; breaches must be answered (STOP → REPORT → WAIT → RECORD). 安全为底 — 可逆放手,不可逆慎行,违约必应(STOP → REPORT → WAIT → RECORD)。 |
| Law | English / 中文 |
|---|---|
| V1 | The Linear Law — an isolated agent creates linear value (1x); value = human-labor substitution × efficiency multiplier; ceiling is visible and decays fast. 线性定律 — 孤立的 Agent 创造线性价值(1x);价值 = 人力替代 × 效率倍数;天花板明显,容易衰减。 |
| V2 | The Leverage Law — an integrated agent creates leveraged value (10x); value = decision quality × decision frequency × impact scope; leveraged but still predictable. 杠杆定律 — 集成的 Agent 创造杠杆价值(10x);价值 = 决策质量 × 决策频次 × 影响范围;杠杆但仍可预测。 |
| V3 | The Emergence Law — coordinated agents create emergent value (100x+); unpredictable, exponential, driven by network effects and emergent behavior. 涌现定律 — 协同的 Agent 创造涌现价值(100x+);不可预测、指数级,由网络效应与涌现行为驱动。 |
Source, protocol contracts, and documentation will be published at github.com/mobfish-ai/cortex-harness-open as the project opens up. 协议合同与文档将陆续发布在 github.com/mobfish-ai/cortex-harness-open。