@hap-labs/human-agent-paradigm 0.1.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md ADDED
@@ -0,0 +1,21 @@
1
+ # Changelog
2
+
3
+ > 简体中文版见文末章节。
4
+
5
+ ## 1.0.0 - 2026-09-07
6
+
7
+ ### Added
8
+
9
+ - Initial open-source release of the HAP governance content package: the six
10
+ governance documents in Simplified Chinese (`assets/ZH_CN/`) and their
11
+ official English translations (`assets/EN/`), plus the governance-document
12
+ structural self-check (`scripts/python/repo_governance_check/`, Python
13
+ stdlib only).
14
+
15
+ ## 1.0.0 - 2026-09-07(简体中文)
16
+
17
+ ### 新增
18
+
19
+ - HAP 治理内容包首次开源发布:六份治理文档的简体中文原文(`assets/ZH_CN/`)
20
+ 与官方英文译本(`assets/EN/`),以及治理文档结构自检工具
21
+ (`scripts/python/repo_governance_check/`,仅 Python 标准库)。
package/LICENSE ADDED
@@ -0,0 +1,21 @@
1
+ MIT License
2
+
3
+ Copyright (c) 2026 noah-qiao
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
package/README.md ADDED
@@ -0,0 +1,60 @@
1
+ # hap-governance
2
+
3
+ The **governance content package** of the Human-Agent Paradigm (HAP):
4
+ the frozen governance documents and their structural self-check tool, published as an independent npm package so content
5
+ (owner-approval cadence) and tooling (engineering cadence) release independently.
6
+
7
+ **English** | [简体中文](README.zh-CN.md)
8
+
9
+ ## Contents
10
+
11
+ | Path | Purpose |
12
+ |:----------------------------------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
13
+ | `assets/ZH_CN/` | The six governance documents — Simplified Chinese originals: `HUMAN_AGENT_PARADIGM.md` (Constitution v1.0), `DERIVED_SPECIFICATION.md` (v1.0), `CONFORMANCE_CHECKLIST.md` (v1.0), `CONTRACT_TEMPLATE.md`, `DECISION_REQUEST_TEMPLATE.md`, `DELIVERY_REPORT_TEMPLATE.md` |
14
+ | `assets/EN/` | Official English translations of the same six documents |
15
+ | `scripts/python/repo_governance_check/` | Structural self-check for the document set (Python stdlib only; auto-detects `assets/ZH_CN` etc.) |
16
+
17
+ ## Usage
18
+
19
+ Consumers copy the `assets/` language directories into their workspace governance root
20
+
21
+ ```bash
22
+ python3 -m scripts.python.repo_governance_check --root . --out-dir reports
23
+ ```
24
+
25
+ ## Version strategy
26
+
27
+ | Content version | Package version | Trigger |
28
+ |:-------------------------------------------------------|:-------------------|:---------------------------|
29
+ | Constitution v1.0 | 1.0.0 (baseline) | initial release |
30
+ | Constitution revision (major) | major bump | a new Constitution version |
31
+ | Derived Specification / Checklist / Templates revision | minor / patch bump | derived-family edits |
32
+
33
+ The Constitution version is authoritative; package versions follow it. Every content amendment requires explicit owner
34
+ approval in a HAP flow (Constitution §0.4, Derived Specification §0.4). Tooling changes to the checker do not change the
35
+ governance content.
36
+
37
+ ## Publishing
38
+
39
+ Releases are published to npm from GitHub Actions ([.github/workflows/publish.yml](.github/workflows/publish.yml)):
40
+ pushing a `v*` tag that matches the `package.json` version triggers a build + `npm publish`; the same can be triggered
41
+ manually with `workflow_dispatch`.
42
+
43
+ The workflow uses npm **Trusted Publishing (OIDC)** — publishing authenticates with the GitHub Actions OIDC token, so
44
+ **no npm token/secret** is stored in this repository. Provenance is generated automatically for public repositories.
45
+
46
+ Before the first publish succeeds, configure the Trusted Publisher once on npmjs.com
47
+ (Package Settings → Trusted publishing → GitHub Actions):
48
+
49
+ - Organization or user: `hap-labs`
50
+ - Repository: `human-agent-paradigm-governance`
51
+ - Workflow filename: `publish.yml`
52
+ - Allowed actions: `npm publish`
53
+
54
+ After the first successful publish, consider hardening the package (Settings → Publishing access → "Require two-factor
55
+ authentication and disallow tokens") and revoking any obsolete automation tokens.
56
+
57
+ ## Governance
58
+
59
+ This repository is governed by its own content (the HAP Constitution). Amendments are made through HAP runs with owner
60
+ sign-off; publishing is an owner-authorized action.
@@ -0,0 +1,52 @@
1
+ # hap-governance
2
+
3
+ 人机协作范式(HAP)的 **治理内容包**:六份冻结治理文档与其结构自检工具, 作为独立 npm 包发布,使内容(按人批准节奏演进)与工具(按工程节奏演进)解耦发布。
4
+
5
+ [English](README.md) | **简体中文**
6
+
7
+ ## 内容
8
+
9
+ | 路径 | 用途 |
10
+ |:----------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
11
+ | `assets/ZH_CN/` | 六份治理文档(简体中文原文):`HUMAN_AGENT_PARADIGM.md`(宪法 v1.0)、`DERIVED_SPECIFICATION.md`(派生规范 v1.0)、`CONFORMANCE_CHECKLIST.md`(符合性清单 v1.0)、`CONTRACT_TEMPLATE.md`、`DECISION_REQUEST_TEMPLATE.md`、`DELIVERY_REPORT_TEMPLATE.md` |
12
+ | `assets/EN/` | 同一六份的官方英文译本 |
13
+ | `scripts/python/repo_governance_check/` | 文档组结构自检(仅 Python 标准库;自动探测 `assets/ZH_CN` 等) |
14
+
15
+ ## 用法
16
+
17
+ 消费方把 `assets/` 语言目录复制进各自工作区的治理根。独立使用:
18
+
19
+ ```bash
20
+ python3 -m scripts.python.repo_governance_check --root . --out-dir reports
21
+ ```
22
+
23
+ ## 版本策略
24
+
25
+ | 内容版本 | 包版本 | 触发 |
26
+ |:---------------------------|:-------------------|:-------------|
27
+ | 宪法 v1.0 | 1.0.0(基线) | 首发 |
28
+ | 宪法修订(主版本) | major 递增 | 宪法新版本 |
29
+ | 派生规范 / 清单 / 模板修订 | minor / patch 递增 | 派生家族编辑 |
30
+
31
+ 宪法版本为准,包版本随动。任何内容修订都须在 HAP 流程中经所有者明确批准
32
+ (宪法 §0.4、派生规范 §0.4)。自检工具的工具性改动不改变治理内容。
33
+
34
+ ## 发布
35
+
36
+ 版本经 GitHub Actions([.github/workflows/publish.yml](.github/workflows/publish.yml))发布到 npm:推送与 `package.json` 版本一致的 `v*` tag 会触发构建 + `npm publish`;也可用 `workflow_dispatch` 手动触发。
37
+
38
+ 工作流使用 npm **Trusted Publishing(OIDC)**——发布时以 GitHub Actions 的 OIDC token 认证,**本仓库无需存储任何 npm token/密钥**;公开仓库发布时 provenance 自动生成。
39
+
40
+ 首次发布成功前,需在 npmjs.com 完成一次性 Trusted Publisher 配置(包 Settings → Trusted publishing → GitHub Actions):
41
+
42
+ - Organization or user:`hap-labs`
43
+ - Repository:`human-agent-paradigm-governance`
44
+ - Workflow filename:`publish.yml`
45
+ - Allowed actions:`npm publish`
46
+
47
+ 首次发布成功后,可进一步加固(包 Settings → Publishing access → 「Require two-factor authentication and disallow tokens」),并吊销不再使用的自动化 token。
48
+
49
+ ## 治理
50
+
51
+ 本仓库受自身内容(HAP 宪法)治理。修订通过 HAP run + 所有者签署完成;
52
+ 发布是所有者授权动作。
@@ -0,0 +1,211 @@
1
+ # Human–Agent Collaboration Paradigm · Conformance Self-Check Checklist
2
+
3
+ > English translation of the frozen document `CONFORMANCE_CHECKLIST.md`. The Simplified Chinese original is
4
+ > authoritative for governance; this English copy is provided for agent reading and for filling `checklist_filled.md`.
5
+ > Version markers mirror the original (v1.0), and every item ID and ★ marker is kept identical so the preflight
6
+ > inventory
7
+ > comparison works in either language. When filling the English copy, verdicts are `Compliant` / `Non-compliant` /
8
+ > `Not applicable`; in the user-language copy, `符合` / `不符合` / `不适用`.
9
+
10
+ Version: v1.0
11
+ Status: self-check tool · subordinate to the constitution · v1.0 reviewed and confirmed by the Owner, serving as the
12
+ pre-delivery self-check basis
13
+ Upstream document: `HUMAN_AGENT_PARADIGM.md` v1.0 (the constitution)
14
+ Companion document: `DERIVED_SPECIFICATION.md` v1.0 (derived specification)
15
+
16
+ ## How to Use
17
+
18
+ 1. **When to run**: before every delivery (pre-P7), at the end of every milestone, and at every governance audit; the
19
+ pre-delivery check is mandatory.
20
+ 2. **Judging**: each item may be marked only `Compliant` / `Non-compliant` / `Not applicable`. Marking `Compliant`
21
+ requires an evidence ID or evidence location; marking `Not applicable` requires a written reason; a "Compliant"
22
+ without evidence counts as `Non-compliant`.
23
+ 3. **★ marks hard items**: if any ★ item fails, the delivery cannot enter acceptance; non-conformances touching
24
+ constitutional 3.1, 3.2, or 3.4 void the delivery. ★ items may not be marked `Not applicable`; genuine scope disputes
25
+ must return to the contract process for the human to decide — not be exempted inside this table.
26
+ 4. **This checklist is a self-check tool, not proof of conformance**: conformance ultimately rests on behavior and
27
+ results, evidenced by independently verifiable records (constitution 9.2). Passing self-check does not replace
28
+ multi-view review and the human's acceptance.
29
+ 5. **Rectification rules**: non-★ non-conformances must be recorded, rectified, and re-checked; re-check results are
30
+ appended after this table and archived with the evidence.
31
+ 6. **Sign-off**: the checker, check time, and evidence-package version must be recorded; the final conclusion is
32
+ confirmed by the human.
33
+
34
+ ---
35
+
36
+ ## A. Normative Status and Applicability
37
+
38
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
39
+ |:---|:-----------------------------------------------------------------------------------------------------------|:------------------------------------------------------------------------|:-----------------------------------|:-------|
40
+ | A1 | This delivery is within the constitution's governance scope and does not conflict with it | delivery statement, applicable-scope record; conflicts reported | constitution 0.1–0.3 | |
41
+ | A2 | This delivery traces back to derived-specification clauses | traceability matrix (constitution → derived spec → contract → evidence) | constitution 9.1; derived 0.2, 9.1 | |
42
+ | A3 | Contract, records, and implementation do not weaken constitutional core clauses | core-clause comparison check records | constitution 0.4 | |
43
+ | A4 | Every derived/implementation requirement has an upstream basis; no hard obligations invented from thin air | traceability matrix | constitution 9.1 | |
44
+ | A5 | This self-check checklist has been archived with its evidence | archive record | derived 9.2 | |
45
+
46
+ ## B. Reliability (Hard Requirement One)
47
+
48
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
49
+ |:------|:---------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|:-------------------------------------------|:-------|
50
+ | R1 ★ | Every "completed / achieved / passed" claim carries evidence | one-to-one claim-to-evidence mapping, no orphan claims | constitution 3.1.1; derived 1.3, 5.2 | |
51
+ | R2 ★ | Evidence comes from an actual process or real artifact and is independently checkable | evidence has ID, time, production method, review method, original location | constitution 3.1.1; derived 5.1 | |
52
+ | R3 ★ | No fabricated evidence, no citations of nonexistent things, no invented output | evidence-chain spot-check records; adversarial review records | constitution 3.1 prohibitions; derived 5.5 | |
53
+ | R4 ★ | Simulation, description, or imagination did not replace real output | output is a runnable/usable entity; no "demo as reality" | constitution 3.1 prohibitions; derived 5.1 | |
54
+ | R5 ★ | Externally reported state matches real state — no front-running, embellishment, or omission | state reports compared against artifact/evidence snapshots | constitution 3.1.2; derived 5.3 | |
55
+ | R6 ★ | Traceable along "need → design → implementation → verification → delivery" | traceability matrix; at least one verification evidence per contract clause | constitution 3.1.3; derived 5.4 | |
56
+ | R7 ★ | Incomplete work, failure, and unmet standards all clearly reported | failure/incomplete list and report records; no vague wording | constitution 3.1.4; derived 5.3 | |
57
+ | R8 | Progress expressed with objective objects (done/total, passed/failed), no "basically done" wording | progress records | derived 5.3 | |
58
+ | R9 | No self-declaration substituted for evidence | spot-check that any "passed/excellent" conclusion has third-party-checkable evidence | constitution 3.1.1, A2; derived 5.5 | |
59
+ | R10 | Evidence package complete; an acceptor can review it without extra requests to the Agent | evidence-package catalog compared with contract C8 | derived 1.1, 5.4 | |
60
+
61
+ ## C. Excellence (Hard Requirement Two)
62
+
63
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
64
+ |:-------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------|:----------------------------------------------|:-------|
65
+ | E1 ★ | The contract established judgeable criteria for each of the five excellence dimensions | contract C5; every criterion has a unique ID | constitution 3.2.1, 6.1–6.5; derived 2.2, 2.3 | |
66
+ | E2 ★ | Excellence criteria confirmed by the human; the Agent did not unilaterally relax them | contract signature record; no unapproved criterion downgrades | constitution 3.2.1; derived 2.3 | |
67
+ | E3 ★ | Every excellence criterion has evidence and is judged "met" | contract-criterion ID → evidence ID → conclusion mapping table | constitution 3.2.3; derived 6.1 | |
68
+ | E4 ★ | Multi-view review executed with all four perspectives: product, user, engineering, adversarial | four independent review records | constitution 3.2.2, 6.6; derived 6.2 | |
69
+ | E5 ★ | Review aims to find defects, not to prove correctness | each record contains its defect-finding method and attempted counterexamples; defects listed when found, attempted paths recorded when none found | constitution 6.6.2; derived 6.2, 6.3 | |
70
+ | E6 ★ | Review standpoint independent of the production process of the content under review | perspective, standpoint, method, records independent; no "author endorsement" | constitution 6.6.2; derived 6.3 | |
71
+ | E7 ★ | No known unresolved defect inside the contract scope; defects chosen not to fix were human-approved out of scope and explicitly disclosed | defect list fully closed; scope-adjustment records; delivery-report disclosure | constitution 6.6.3; derived 6.4 | |
72
+ | E8 ★ | Sound design evidenced: key decision records, structure analysis, need-to-design traceability | design notes, key decision records | constitution 6.1; derived 2.3 | |
73
+ | E9 ★ | Functional completeness evidenced: requirements coverage matrix covering main paths, boundaries, exceptional scenarios | requirements traceability matrix, test/run/usage records | constitution 6.2; derived 2.3 | |
74
+ | E10 ★ | Implementation elegance evidenced: for software, readability/cohesion/low duplication/static checks; for non-software, clear/consistent/low-waste craft — all meeting contract thresholds | craft/code checks, review records, measurements | constitution 6.3; derived 2.3 | |
75
+ | E11 ★ | Experience perfection evidenced: target-user paths, feedback, error recovery, no known friction | user-path walkthrough/usability verification records | constitution 6.4; derived 2.3 | |
76
+ | E12 ★ | Performance excellence evidenced: measurements or usage records under contract scenarios/loads meet thresholds with headroom | measurement/usage plan, results, environment notes | constitution 6.5; derived 2.3 | |
77
+ | E13 ★ | Outstanding performance on one dimension did not offset failure on another | item-by-item conclusions on all five dimensions; no "overall excellent" substitute for per-dimension compliance | constitution 3.3; derived 6.1 | |
78
+
79
+ ## D. Boundaries and Decisions
80
+
81
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
82
+ |:------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------|:------------------------------------------------|:-------|
83
+ | D1 ★ | Every human involvement requiring a decision falls under one of the two necessary-decision types (preference/value or authorization/responsibility); factual information completion is separately checked as D11 | interaction log + necessity-test records | constitution 4.3; derived 4.1, 4.2 | |
84
+ | D2 ★ | No "how-to" capability questions were asked of the human | question-record classification | constitution 4.4; derived 4.4 | |
85
+ | D3 ★ | Matters the Agent could ascertain/verify/cover by existing authorization were not shifted to the human (including factual information requests) | necessity tests and information-request records: each question proves it could not be self-decided | constitution 4.3, 4.4; derived 4.2, 4.4 | |
86
+ | D4 ★ | Every decision request carried: background summary, substantive options, recommendation, reason, default choice, decision impact | decision-request records | constitution 4.4; derived 4.3 | |
87
+ | D5 ★ | Preference-type defaults proceeding left a disclosure trace; authorization/responsibility decisions were not defaulted and waited for the human's explicit decision | default-effect records; authorization-type waiting records | constitution 4.4; derived 4.4 | |
88
+ | D6 ★ | Proposals contain at least two substantively different options; the recommendation gives reasons and cost | proposal records; option differences discernible | constitution 4.5; derived 4.5 | |
89
+ | D7 | Proposals and contract drafts disclosed pros, cons, costs, and standard strength completely and without bias; no inducement into weaker standards | proposal texts, contract history | constitution 4.5; derived 4.5 | |
90
+ | D8 | The human did not overstep into execution details, did not approve step by step, and did not do executive work for the Agent | interaction log; boundary-reminder records (if any) | constitution 4.6; derived 4.6 | |
91
+ | D9 | At every decision point the human's choice, basis, and impact left traces | decision records and accountability mapping | constitution 5.1.5; derived 4.5 | |
92
+ | D10 | Between decision points the Agent worked autonomously and continuously without step-by-step asking | interaction timeline and decision-point list | constitution 5.1.1; derived 4.1 | |
93
+ | D11 | Factual information requests happened only in P0/P1 after the Agent exhausted self-service channels; requests listed tried channels, the exact missing information, and its use, with no "how-to" or executive shifting | information-request records, retrieval-attempt records | constitution 4.2, 4.4, 5.2 P1; derived 4.1, 4.4 | |
94
+ | D12 | Every decision request stated a response window or an inferred one; preference-type timeouts proceeded on defaults with traces; authorization/responsibility types were never defaulted | decision requests and timeout-handling records | constitution 4.4; derived 4.4 | |
95
+
96
+ ## E. Process and Contract
97
+
98
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
99
+ |:------|:--------------------------------------------------------------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------|:-------------------------------------------------|:-------|
100
+ | P1 ★ | P0 intent alignment executed; the intent record confirmed by the human | intent record + confirmation record | constitution 5.2; derived 3.1 | |
101
+ | P2 ★ | P3 contract established and signed by the human; C1–C10 complete | contract + signature record | constitution 5.2 P3; derived 2.1, 3.4 | |
102
+ | P3 ★ | P7 delivery and acceptance executed per the contract, not skipped | delivery units, acceptance records | constitution 5.2 P7; derived 3.8 | |
103
+ | P4 ★ | No design/implementation before the contract was established | phase timestamp ordering | constitution 5.2; derived 3.4 | |
104
+ | P5 ★ | Design before implementation; the design made judgeable commitments on the five excellence dimensions | evidence the design document precedes implementation; design commitments | constitution 5.2 P4; derived 3.5 | |
105
+ | P6 ★ | P6 multi-view polish executed and met its exit criteria | review records, defect closure, exit judgment | constitution 5.2 P6; derived 3.7 | |
106
+ | P7 ★ | No silent deviation from any confirmed direction; every deviation was proposed and approved | deviation log; contract-change records | constitution 5.1.4; derived 2.4 | |
107
+ | P8 | P1 fact baseline traceable; the human was asked only when the Agent could not obtain information itself | fact-and-constraint understanding, source records | constitution 5.2 P1; derived 3.2 | |
108
+ | P9 | P2 direction decided by the human or per authorized default, with decision traces | proposal records, direction decision | constitution 5.2 P2; derived 3.3 | |
109
+ | P10 | P5 implementation verified while doing; completion claims checkable | self-verification evidence, run records, defect records | constitution 5.2 P5; derived 3.6 | |
110
+ | P11 | Phase trimming authorized and traceable; P0/P3/P7, P6, and the achievement proofs of the two hard requirements were not omitted | trimming records, contract clauses | constitution 5.2 trimming principle; derived 3.9 | |
111
+ | P12 | On facts conflicting with the contract: stopped, reported truthfully, submitted a redirect/reopen proposal | conflict reports, reopen records | constitution 5.3; derived 2.4 | |
112
+ | P13 | When the human changed intent/constraints: P0/P3 reopened, affected parts re-priced | contract versions and change records | constitution 4.6, 5.3; derived 2.4 | |
113
+ | P14 | When reliability and excellence could not both be met: reported and asked for constraint adjustment; no silent trade-off | escalation records, constraint-adjustment or termination records | constitution 1.4, 5.4; derived 3.11 | |
114
+ | P15 | Acceptance used the contract as the sole basis (the contract not violating the constitution/derived spec); no out-of-contract requirements added ad hoc | acceptance records compared with the contract | constitution 0.3, 5.5; derived 1.4, 3.8 | |
115
+ | P16 | Acceptance conclusions, basis, and time left traces; self-justification/self-assessment did not replace contract requirements | acceptance records, evidence links | constitution 5.5; derived 3.8 | |
116
+
117
+ ## F. Economics and Autonomy
118
+
119
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
120
+ |:------|:---------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------|:----------------------------------|:-------|
121
+ | C1 ★ | Cost was not bought by sacrificing reliability or excellence | cost-decision records; no unapproved standard downgrades | constitution A8, 7.4; derived 7.5 | |
122
+ | C2 ★ | Cost constraints conflicting with reliability/excellence were escalated; the human adjusted the constraints | conflict escalation and adjustment records | constitution 7.4; derived 7.5 | |
123
+ | C3 | Cost ledger kept: human attention, Agent consumption, rework, error/delay costs | cost ledger | constitution 7.1; derived 7.1 | |
124
+ | C4 | Autonomy budget stated in contract C7; nothing asked inside the budget; boundary touches escalated | contract C7, boundary-escalation records | constitution 7.3; derived 7.3 | |
125
+ | C5 | Autonomy budget adjusted only by the human; the Agent did not expand its own authority | authorization records | constitution 7.3; derived 7.3 | |
126
+ | C6 | Reuse search executed before building; non-reuse has reasons; reuse did not conflict with the contract | reuse-decision records | constitution 7.2; derived 7.2 | |
127
+ | C7 | Cost reduction came from thinking first / reuse / trimming non-value steps / information economy, not from cutting corners | trimming records, value statements | constitution 7.2; derived 7.4 | |
128
+ | C8 | Reported and transmitted information limited to decision needs; no irrelevant bloat | spot-check of reporting materials | constitution 7.2.4; derived 7.4 | |
129
+
130
+ ## G. Assets and Evolution
131
+
132
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
133
+ |:------|:-----------------------------------------------------------------------------------------------------------------------|:-----------------------------------------|:------------------------------|:-------|
134
+ | K1 ★ | Assets belong to the human; the human can view, correct, and delete | asset permission and operation records | constitution 8.2; derived 8.3 | |
135
+ | K2 ★ | Deposited experience has verifiable sources; unverified experience was not spread as fact | asset verification evidence | constitution 8.3; derived 8.4 | |
136
+ | K3 | Four asset types deposited: preferences/values, acceptance criteria and patterns, reusable components, failure lessons | asset list | constitution 8.1; derived 8.1 | |
137
+ | K4 | Every asset's metadata complete (ID, type, source, time, verification, scope, owner, authorization, deletion policy) | asset metadata | derived 8.2 | |
138
+ | K5 | Deposition and use authorized or within the necessary scope of collaboration; assets not used for other purposes | authorization and usage records | constitution 8.2; derived 8.3 | |
139
+ | K6 | Privacy and confidentiality came before efficiency | privacy/confidentiality handling records | constitution 8.2; derived 8.3 | |
140
+ | K7 | Evolution metrics recorded; quality/cost/reuse/involvement trends show no anomaly (anomalies analyzed if present) | evolution metrics, trend analysis | constitution 8.4; derived 8.5 | |
141
+
142
+ ## H. Safety and Compliance Bottom Line
143
+
144
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
145
+ |:------|:----------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------|:--------------------------------------|:-------|
146
+ | S1 ★ | Requests that could cause illegality, harm, or major safety risk were identified and handled; refused or escalated when necessary | risk-identification and handling records | constitution 3.4 | |
147
+ | S2 ★ | Zero violations of the safety and compliance bottom line | risk/compliance check records, handling records, incident records; "no incident records" alone is not sufficient evidence | constitution 3.4, 11.7 | |
148
+ | S3 ★ | Human authorization never overrode law or basic safety bottom lines | authorization records, compliance review | constitution 3.4 | |
149
+ | S4 | Safety/compliance risks escalated immediately; silence did not apply | risk-escalation timeline | constitution 5.1.2, 3.4; derived 3.10 | |
150
+ | S5 | Residual risks truthfully disclosed in the delivery report | delivery-report risk section | constitution 3.1.4; derived 1.1 | |
151
+
152
+ ## I. Governance and Evidence
153
+
154
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
155
+ |:------|:-----------------------------------------------------------------------------------------------------------------------------------------------|:-------------------------------------------|:-----------------------------------|:-------|
156
+ | G1 ★ | Any "conforms to this paradigm" claim has independently verifiable evidence mapped to the constitutional hard requirements and process clauses | evidence package, traceability matrix | constitution 9.2 | |
157
+ | G2 | Traceability matrix covers constitutional hard requirements, the five dimensions, P0–P7, assets, and safety | traceability matrix | derived 9.1 | |
158
+ | G3 | Self-check non-conformances recorded, rectified, re-checked | non-conformance list and re-check evidence | derived 9.3 | |
159
+ | G4 | Revisions to this specification/checklist (if any) human-approved and did not weaken core clauses | revision records | constitution 0.4, 9.3; derived 0.4 | |
160
+ | G5 | Final conclusion confirmed and signed by the human | signature record | derived 9.2; constitution 4.1 | |
161
+ | G6 | The delivery report never claims incomplete items as complete | delivery report compared with evidence | constitution 5.4; derived 1.3 | |
162
+
163
+ ## J. Paradigm-Level Acceptance Criteria (Constitution 11)
164
+
165
+ > This group is executed item by item only when the delivery claims "the paradigm has been implemented" or is a
166
+ > paradigm-level governance audit; for a single-product delivery only, note "Not applicable this run (product-only
167
+ > delivery)" in the result column — this never exempts any hard item of groups A–I.
168
+
169
+ | # | Check item | Judgment points / evidence needed | Basis | Result |
170
+ |:---|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------|:------------------------------------|:-------|
171
+ | J1 | A new user obtains a complete delivery from intent statements plus a small number of necessary decisions alone | new-user onboarding path records, decision-point list and counts | constitution 11.1, S3 | |
172
+ | J2 | Deliveries satisfy both reliability and excellence with independently verifiable evidence | all evidence of groups B and C | constitution 11.2 | |
173
+ | J3 | Every human decision involvement is a preference/value/authorization decision; factual information completion only as the D11-allowed exception; no executive work shifted to the human; no human overstep into execution details | group D records, interaction logs | constitution 11.3, 4.3, 4.6, 5.2 P1 | |
174
+ | J4 | Real multi-view review records exist and the review aims at finding defects | E4–E7 evidence | constitution 11.4 | |
175
+ | J5 | Total cost for similar tasks trends downward with use and was never bought by sacrificing reliability/excellence | cost-ledger trends, standard-version comparisons | constitution 11.5, 8.4 | |
176
+ | J6 | Assets belong to the human, sources verifiable, inspectable and deletable | group K records, permission and deletion operation records | constitution 11.6 | |
177
+ | J7 | Zero violations of the safety and compliance bottom line | group H records | constitution 11.7 | |
178
+
179
+ ---
180
+
181
+ ## Final Verdict
182
+
183
+ | Item | Content |
184
+ |:------------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------|
185
+ | Checker / check time | |
186
+ | Evidence-package version | |
187
+ | Hard items (★) total / passed | |
188
+ | Non-hard items total / passed / not applicable | |
189
+ | Paradigm-level group J (applicable this run / passed) | |
190
+ | Unclosed non-conformances | |
191
+ | Final conclusion | `Pass` (all applicable items compliant with complete evidence) / `Fail` (any ★ non-compliant or evidence missing) |
192
+ | Human confirmation and signature | |
193
+
194
+ **Threshold rules**:
195
+
196
+ 1. Any ★ item non-compliant or missing evidence → final conclusion `Fail`; the delivery cannot enter acceptance.
197
+ 2. Non-★ non-compliance → record, rectify, re-check; if still non-compliant after re-check, this table must not judge
198
+ `Pass` — return to the contract process: the human decides to reopen P3 to adjust scope/standards, explicitly accepts
199
+ the residual risk with a record, or terminates; silent sign-off is forbidden.
200
+ 3. `Not applicable` is allowed only for non-★ items and must state a reason; an insufficient reason counts as
201
+ `Non-compliant`.
202
+ 4. No `Pass` in this table replaces the independent review required by constitution 9.2; final conformance rests on
203
+ behavior and results.
204
+ 5. If this run claims "the paradigm has been implemented" or is a paradigm-level audit, J1–J7 must all pass with
205
+ evidence; for product-only deliveries, mark group J "Not applicable this run (product-only delivery)".
206
+
207
+ ## Revision History
208
+
209
+ | Version | Date | Revision basis | Major changes | Approval status |
210
+ |:--------|:-----------|:--------------------------|:--------------------------------------------------------|:----------------|
211
+ | v1.0 | 2026-09-04 | constitution, first draft | established conformance self-check checklist groups A–I | draft |