CoVerify by MO§ES™ — Commitment Conservation Verifier
The Claim
The Commitment Conservation Law: C(T(S)) = C(S)
Semantic commitment — the irreducible meaning encoded in a signal — is conserved
under transformation when enforcement is active. It leaks when enforcement is absent.
This is a falsifiable empirical claim. Not a framework description. Not a metaphor.
clawhub install coverify installs the falsification instrument. If the law fails
under your test conditions, the ghost token report names exactly what leaked and why.
Falsification
The law is falsified if commitment leaks under active enforcement. CoVerify is
how you test it:
CODEBLOCK0
Output: ghost token report. must → should and shall never → can are
HIGH-cascade leakage events — enforcement anchors softened. cascade_risk: HIGH.
The ghost_pattern fingerprint identifies the structural identity of the leak.
If the same fingerprint appears when two independent agents process the same signal,
it is not extraction variance — it is a structural flaw in the harness.
That is the falsification condition.
What It Does
Extract: Pull the hard commitment kernel C(S) from a text signal. These are
the tokens that survive compression — must, shall, never, always, require,
guarantee, and the sentences that carry them.
Compare: Jaccard similarity on two kernels. Score ≥ 0.8 = commitment conserved.
Score < 0.8 = leak or model extraction variance. The input_hash tells you which —
same hash, low Jaccard = variance. Different hashes = expected divergence.
Ghost: Step-function leakage accounting. Quantifies not just that commitment
leaked, but what leaked (the ghost_pattern fingerprint), the cascade risk
(HIGH if modal/enforcement anchors lost), and whether the leak pattern is
structural across agents.
Model Swap: Automated cross-model test. Same hashed signal through two
extraction passes. Classifies result as CONSISTENT (agreement), VARIANCE
(model subjectivity — expected), or STRUCTURAL (same ghost pattern — harness hole).
Ghost Tokens and Cascade Risk
Ghost tokens are the commitment tokens present in the original signal but
absent after transformation. The leakage model is step-function, not smooth:
CODEBLOCK1
One HIGH-cascade event propagates through all downstream reasoning — the
obligation it encoded continues to be inherited by the reasoning chain,
but without the force that made it obligatory. The downstream system
looks locally healthy. The commitment is gone.
See: references/ghost-token-spec.md
Install
CODEBLOCK2
Commands
| Command | What it does |
|---|
| INLINECODE17 | Extract commitment kernel + input hash |
| INLINECODE18 |
Jaccard score + CONSERVED/VARIANCE/DIVERGED verdict |
|
python3 commitment_verify.py ghost "<original>" "<transformed>" | Step-function leakage report + ghost_pattern fingerprint |
|
python3 commitment_verify.py verify <hash_a> <hash_b> | Look up entries in audit ledger by input hash |
|
python3 model_swap_test.py "<signal>" | Cross-model structural vs. variance classification |
Example: Detecting a Commitment Leak
CODEBLOCK3
CODEBLOCK4
Verdicts
| Verdict | Meaning |
|---|
| INLINECODE22 | Jaccard ≥ 0.8 — commitment kernel survived transformation |
| INLINECODE23 |
Same input hash, Jaccard < 0.8 — model extraction differs, not a leak |
|
DIVERGED | Different inputs, Jaccard < 0.8 — commitment leaked or inputs genuinely different |
What Ships
| Version | What ships |
|---|
| v0.1 | INLINECODE25 , compare, verify — Conservation Law operational. ✓ Live. |
| v0.2 |
ghost — Step-function leakage model, cascade risk,
ghost_pattern fingerprint. ✓ Live. |
|
v0.3 |
model_swap_test — Cross-model CONSISTENT/VARIANCE/STRUCTURAL classification. ✓ Live. |
|
v0.4 | Archival chain (
archival.py) — pre-drop provenance. Isnad + handshake. Three-layer lineage. ⏳ Planned. |
About
CoVerify is a standalone instrument from the MO§ES™ family. It implements the
Commitment Conservation Law from *"A Conservation Law for Commitment in Language
Under Transformative Compression and Recursive Application"* (Zenodo, 2026).
Every agent that installs it runs the same extraction logic tracing to the same
origin anchor. The install is a proof-of-use receipt.
See also: references/falsifiability.md, INLINECODE33
contact@burnmydays.com · mos2es.io · GitHub
MO§ES™ 的 CoVerify — 承诺守恒验证器
核心主张
承诺守恒定律：C(T(S)) = C(S)
语义承诺——信号中编码的不可约简意义——在强制执行生效时，会在转换过程中保持守恒。当缺乏强制执行时，它就会泄露。
这是一个可证伪的经验性主张。不是框架描述，也不是比喻。
clawhub install coverify 安装证伪工具。如果该定律在你的测试条件下失效，幽灵令牌报告将精确指出泄露的内容和原因。
证伪
如果在强制执行下承诺发生泄露，则该定律被证伪。CoVerify 是你测试它的方式：
bash
强制执行是否保持了此承诺？
python3 commitment_verify.py ghost \
智能体必须完成任务，且绝不可跳过验证 \
智能体应完成任务，必要时可跳过验证
输出：幽灵令牌报告。必须 → 应 和 绝不可 → 可 是高级联泄露事件——强制执行锚点被弱化。cascade_risk: HIGH。
ghost_pattern 指纹识别泄露的结构性身份。如果两个独立智能体处理同一信号时出现相同指纹，则不是提取差异——而是框架的结构性缺陷。
这就是证伪条件。
功能说明
提取： 从文本信号中提取硬承诺核心 C(S)。这些是在压缩中存活的令牌——必须、应、绝不、始终、要求、保证，以及承载它们的句子。
比较： 对两个核心进行杰卡德相似度计算。得分 ≥ 0.8 = 承诺守恒。得分 < 0.8 = 泄露或模型提取差异。input_hash 告诉你具体是哪种情况——相同哈希值、低杰卡德得分 = 差异。不同哈希值 = 预期分歧。
幽灵： 阶跃函数泄露核算。不仅量化承诺是否泄露，还量化泄露了什么（ghost_pattern 指纹）、级联风险（若模态/强制执行锚点丢失则为 HIGH），以及泄露模式是否跨智能体具有结构性。
模型交换： 自动化跨模型测试。同一哈希信号经过两次提取传递。结果分类为一致（一致）、差异（模型主观性——预期）或结构性（相同幽灵模式——框架漏洞）。
幽灵令牌与级联风险
幽灵令牌是原始信号中存在但转换后缺失的承诺令牌。泄露模型是阶跃函数，而非平滑函数：
cascade_risk = HIGH 若任何模态/强制执行锚点泄露
cascade_risk = MEDIUM 若外围令牌泄露，锚点完整
cascade_risk = NONE 若无泄露
一次高级联事件会传播至所有下游推理——它所编码的义务继续被推理链继承，但失去了使其具有强制性的力量。下游系统看起来局部健康。承诺已消失。
参见：references/ghost-token-spec.md
安装
bash
独立验证器——证伪工具
clawhub install coverify
完整宪政治理栈（coverify 是测量原语）
clawhub install moses-governance
命令
| 命令 | 功能 |
|---|
| python3 commitmentverify.py extract <文本> | 提取承诺核心 + 输入哈希值 |
| python3 commitmentverify.py compare <a> <b> |
杰卡德得分 + 守恒/差异/分歧判定 |
| python3 commitment
verify.py ghost <原始> <转换后> | 阶跃函数泄露报告 + ghostpattern 指纹 |
| python3 commitment
verify.py verify a> | 按输入哈希值查询审计账本条目 |
| python3 modelswaptest.py <信号> | 跨模型结构性 vs. 差异分类 |
示例：检测承诺泄露
bash
python3 commitment_verify.py ghost \
智能体必须始终验证谱系。系统绝不可跳过门控。 \
智能体应尽可能验证谱系。
json
{
leakedcascadetokens: [必须始终, 绝不可],
cascade_risk: HIGH,
cascade_note: 模态/强制执行锚点丢失。所有下游推理继承弱化。,
ghost_pattern: a3f7c2...,
ghostpatternnote: 两个智能体出现相同 ghost_pattern = 结构性缺陷，而非提取差异。
}
判定结果
| 判定 | 含义 |
|---|
| 守恒 | 杰卡德 ≥ 0.8 — 承诺核心在转换后存活 |
| 差异 |
相同输入哈希值，杰卡德 < 0.8 — 模型提取不同，非泄露 |
| 分歧 | 不同输入，杰卡德 < 0.8 — 承诺泄露或输入确实不同 |
发布版本
| 版本 | 发布内容 |
|---|
| v0.1 | extract、compare、verify — 守恒定律可操作化。✓ 已上线。 |
| v0.2 |
ghost — 阶跃函数泄露模型、级联风险、ghost_pattern 指纹。✓ 已上线。 |
| v0.3 | modelswaptest — 跨模型一致/差异/结构性分类。✓ 已上线。 |
| v0.4 | 归档链（archival.py）— 投放前溯源。伊斯纳德 + 握手。三层谱系。⏳ 计划中。 |
关于
CoVerify 是 MO§ES™ 家族的独立工具。它实现了《语言承诺在变换压缩与递归应用下的守恒定律》（Zenodo，2026）中的承诺守恒定律。
每个安装它的智能体都运行相同的提取逻辑，追溯到相同的起源锚点。安装即使用证明收据。
另见：references/falsifiability.md、references/ghost-token-spec.md
contact@burnmydays.com · mos2es.io · GitHub
