@sakanano-moe/agenthelm 6.13.1 → 6.14.0

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.
Files changed (117) hide show
  1. package/package.json +1 -1
  2. package/runtime/index.json +6 -6
  3. package/runtime/plugins/agent-pack/.codex-plugin/plugin.json +1 -1
  4. package/runtime/plugins/agent-pack/plugin.json +1 -1
  5. package/runtime/plugins/agent-pack/references/capability-config.md +10 -0
  6. package/runtime/plugins/agent-pack/references/check-opinion-migrate.en.md +31 -0
  7. package/runtime/plugins/agent-pack/references/check-opinion-migrate.md +31 -0
  8. package/runtime/plugins/agent-pack/references/dedupe-contract.en.md +27 -0
  9. package/runtime/plugins/agent-pack/references/dedupe-contract.md +27 -0
  10. package/runtime/plugins/agent-pack/references/inventory-contract.en.md +83 -0
  11. package/runtime/plugins/agent-pack/references/inventory-contract.md +83 -0
  12. package/runtime/plugins/agent-pack/references/nudge.en.md +10 -2
  13. package/runtime/plugins/agent-pack/references/nudge.md +10 -2
  14. package/runtime/plugins/agent-pack/references/relocate-contract.en.md +41 -0
  15. package/runtime/plugins/agent-pack/references/relocate-contract.md +41 -0
  16. package/runtime/plugins/agent-pack/runtime/agent_pack_claude.py +572 -30
  17. package/runtime/plugins/agent-pack/runtime/agent_pack_clients.py +233 -0
  18. package/runtime/plugins/agent-pack/runtime/agent_pack_config.py +112 -4
  19. package/runtime/plugins/agent-pack/runtime/agent_pack_dedupe.py +502 -0
  20. package/runtime/plugins/agent-pack/runtime/agent_pack_desktop.py +92 -0
  21. package/runtime/plugins/agent-pack/runtime/agent_pack_ext_common.py +251 -0
  22. package/runtime/plugins/agent-pack/runtime/agent_pack_inventory.py +989 -0
  23. package/runtime/plugins/agent-pack/runtime/agent_pack_mcp_write.py +447 -0
  24. package/runtime/plugins/agent-pack/runtime/agent_pack_nudge.py +29 -4
  25. package/runtime/plugins/agent-pack/runtime/agent_pack_relocate.py +385 -0
  26. package/runtime/plugins/agent-pack/runtime/agent_pack_scope.py +233 -0
  27. package/runtime/plugins/agent-pack/runtime/agenthelm.py +183 -14
  28. package/runtime/plugins/agent-pack/runtime/locale/en.json +250 -1
  29. package/runtime/plugins/agent-pack/runtime/mcp_server.py +12 -0
  30. package/runtime/plugins/agent-pack/runtime/nudge_text.py +8 -2
  31. package/runtime/plugins/agent-pack/runtime/user_settings.py +2 -1
  32. package/runtime/plugins/agent-pack/skills/agent-pack/SKILL.md +15 -0
  33. package/runtime/plugins/agents-init/.codex-plugin/plugin.json +1 -1
  34. package/runtime/plugins/agents-init/content.json +4 -2
  35. package/runtime/plugins/agents-init/plugin.json +1 -1
  36. package/runtime/plugins/agents-init/skills/agents-init/SKILL.md +2 -0
  37. package/runtime/plugins/agents-init/skills/agents-init/launcher.en.sh +93 -0
  38. package/runtime/plugins/agents-init/skills/agents-init/project.schema.json +27 -0
  39. package/runtime/plugins/agents-init/skills/agents-init/scripts/baseline.py +4 -1
  40. package/runtime/plugins/agents-init/skills/agents-init/scripts/init_project.py +24 -9
  41. package/runtime/plugins/agents-init/skills/agents-init/skill.json +1 -1
  42. package/runtime/plugins/opinion-manager/.codex-plugin/plugin.json +1 -1
  43. package/runtime/plugins/opinion-manager/catalog/README.md +3 -1
  44. package/runtime/plugins/opinion-manager/catalog/rules/om.align-requirements/versions/1.1.0.json +34 -0
  45. package/runtime/plugins/opinion-manager/catalog/rules/om.assumptions-authority/versions/1.1.0.json +34 -0
  46. package/runtime/plugins/opinion-manager/catalog/rules/om.batch-questions/versions/1.1.0.json +34 -0
  47. package/runtime/plugins/opinion-manager/catalog/rules/om.bilingual-terms/versions/1.1.0.json +34 -0
  48. package/runtime/plugins/opinion-manager/catalog/rules/om.calculations/versions/1.1.0.json +34 -0
  49. package/runtime/plugins/opinion-manager/catalog/rules/om.chinese-response/versions/1.1.0.json +34 -0
  50. package/runtime/plugins/opinion-manager/catalog/rules/om.clean-deliverable/versions/1.1.0.json +34 -0
  51. package/runtime/plugins/opinion-manager/catalog/rules/om.complete-options/versions/1.1.0.json +34 -0
  52. package/runtime/plugins/opinion-manager/catalog/rules/om.conclusion-first/versions/1.1.0.json +34 -0
  53. package/runtime/plugins/opinion-manager/catalog/rules/om.concrete-titles/versions/1.1.0.json +34 -0
  54. package/runtime/plugins/opinion-manager/catalog/rules/om.consistent-items/versions/1.1.0.json +34 -0
  55. package/runtime/plugins/opinion-manager/catalog/rules/om.device-readable/versions/1.1.0.json +34 -0
  56. package/runtime/plugins/opinion-manager/catalog/rules/om.editable-diagrams/versions/1.1.0.json +34 -0
  57. package/runtime/plugins/opinion-manager/catalog/rules/om.evidence-inference/versions/1.1.0.json +34 -0
  58. package/runtime/plugins/opinion-manager/catalog/rules/om.maintenance-chinese/versions/1.1.0.json +34 -0
  59. package/runtime/plugins/opinion-manager/catalog/rules/om.maintenance-project-language/versions/1.1.0.json +34 -0
  60. package/runtime/plugins/opinion-manager/catalog/rules/om.markdown-structure/versions/1.1.0.json +34 -0
  61. package/runtime/plugins/opinion-manager/catalog/rules/om.portable-files/versions/1.1.0.json +34 -0
  62. package/runtime/plugins/opinion-manager/catalog/rules/om.report-context/versions/1.1.0.json +34 -0
  63. package/runtime/plugins/opinion-manager/catalog/rules/om.report-formats/versions/1.1.0.json +34 -0
  64. package/runtime/plugins/opinion-manager/catalog/rules/om.report-review/versions/1.1.0.json +34 -0
  65. package/runtime/plugins/opinion-manager/catalog/rules/om.response-brief/versions/1.1.0.json +34 -0
  66. package/runtime/plugins/opinion-manager/catalog/rules/om.response-dense/versions/1.1.0.json +34 -0
  67. package/runtime/plugins/opinion-manager/catalog/rules/om.rules-tools/versions/1.1.0.json +34 -0
  68. package/runtime/plugins/opinion-manager/catalog/rules/om.semantic-wrap/versions/1.1.0.json +34 -0
  69. package/runtime/plugins/opinion-manager/catalog/rules/om.solutions-proportional/versions/1.1.0.json +34 -0
  70. package/runtime/plugins/opinion-manager/catalog/rules/om.solutions-three-level/versions/1.1.0.json +34 -0
  71. package/runtime/plugins/opinion-manager/catalog/rules/om.tables-compact/versions/1.1.0.json +34 -0
  72. package/runtime/plugins/opinion-manager/catalog/rules/om.tables-list-first/versions/1.1.0.json +34 -0
  73. package/runtime/plugins/opinion-manager/catalog/rules/om.test-evidence/versions/1.1.0.json +34 -0
  74. package/runtime/plugins/opinion-manager/catalog/rules/om.versioned-files/versions/1.1.0.json +34 -0
  75. package/runtime/plugins/opinion-manager/catalog/templates/delivery-standard/compact-tables/versions/1.1.0.json +32 -0
  76. package/runtime/plugins/opinion-manager/catalog/templates/delivery-standard/list-first/versions/1.1.0.json +32 -0
  77. package/runtime/plugins/opinion-manager/catalog/templates/development-standard/chinese-maintenance/versions/1.1.0.json +26 -0
  78. package/runtime/plugins/opinion-manager/catalog/templates/development-standard/project-language/versions/1.1.0.json +26 -0
  79. package/runtime/plugins/opinion-manager/catalog/templates/response-style/brief-chinese/versions/1.1.0.json +30 -0
  80. package/runtime/plugins/opinion-manager/catalog/templates/response-style/dense-chinese/versions/1.1.0.json +30 -0
  81. package/runtime/plugins/opinion-manager/catalog/templates/scenario/concise-project/versions/1.1.0.json +181 -0
  82. package/runtime/plugins/opinion-manager/catalog/templates/scenario/reviewed-project/versions/1.1.0.json +179 -0
  83. package/runtime/plugins/opinion-manager/catalog/templates/work-method/proportional/versions/1.1.0.json +31 -0
  84. package/runtime/plugins/opinion-manager/catalog/templates/work-method/three-level/versions/1.1.0.json +31 -0
  85. package/runtime/plugins/opinion-manager/content.json +44 -3
  86. package/runtime/plugins/opinion-manager/plugin.json +1 -1
  87. package/runtime/plugins/opinion-manager/references/inbox.en.md +25 -4
  88. package/runtime/plugins/opinion-manager/references/inbox.md +25 -4
  89. package/runtime/plugins/opinion-manager/references/profile.schema.json +19 -2
  90. package/runtime/plugins/opinion-manager/references/rule.schema.json +25 -3
  91. package/runtime/plugins/opinion-manager/references/versioned-template.schema.json +5 -4
  92. package/runtime/plugins/opinion-manager/references/versioning.md +89 -4
  93. package/runtime/plugins/opinion-manager/runtime/context.py +26 -19
  94. package/runtime/plugins/opinion-manager/runtime/inbox.py +400 -25
  95. package/runtime/plugins/opinion-manager/runtime/inbox_relations.py +106 -0
  96. package/runtime/plugins/opinion-manager/runtime/inbox_routes.py +132 -0
  97. package/runtime/plugins/opinion-manager/runtime/inbox_text.py +117 -11
  98. package/runtime/plugins/opinion-manager/runtime/lifecycle.py +35 -5
  99. package/runtime/plugins/opinion-manager/runtime/locale/en.json +137 -2
  100. package/runtime/plugins/opinion-manager/runtime/migrate.py +582 -0
  101. package/runtime/plugins/opinion-manager/runtime/opinion_manager.py +9 -4
  102. package/runtime/plugins/opinion-manager/runtime/rule_lines.py +247 -0
  103. package/runtime/plugins/opinion-manager/runtime/tags.py +217 -0
  104. package/runtime/plugins/opinion-manager/runtime/versioned.py +289 -41
  105. package/runtime/plugins/opinion-manager/skills/opinion-manager/SKILL.md +8 -7
  106. package/runtime/plugins/project-orchestrator/.codex-plugin/plugin.json +1 -1
  107. package/runtime/plugins/project-orchestrator/plugin.json +1 -1
  108. package/runtime/plugins/project-orchestrator/runtime/locale/en.json +305 -3
  109. package/runtime/plugins/project-orchestrator/runtime/private_paths.py +11 -1
  110. package/runtime/plugins/project-orchestrator/runtime/project_cli.py +101 -1
  111. package/runtime/plugins/project-orchestrator/runtime/project_demo.py +13 -1
  112. package/runtime/plugins/project-orchestrator/runtime/project_ext_actions.py +415 -0
  113. package/runtime/plugins/project-orchestrator/runtime/project_extensions.py +971 -0
  114. package/runtime/plugins/project-orchestrator/runtime/project_extensions_demo.py +86 -0
  115. package/runtime/plugins/project-orchestrator/runtime/project_menu.py +345 -2
  116. package/runtime/plugins/project-orchestrator/runtime/project_real.py +17 -1
  117. package/runtime/plugins/project-orchestrator/skills/project-orchestrator/SKILL.md +2 -0
@@ -0,0 +1,181 @@
1
+ {
2
+ "family": "scenario",
3
+ "variant": "concise-project",
4
+ "version": "1.1.0",
5
+ "name": {
6
+ "zh": "简明沟通与项目交付组合",
7
+ "en": "Concise communication with project delivery"
8
+ },
9
+ "description": {
10
+ "zh": "用途:需要简明沟通、按需分析和项目语言维护的任务。 不适用:已要求完整固定分析框架的任务。 这是可选择的行为偏好,不代表普遍最优做法。",
11
+ "en": "Use: tasks needing concise communication, on-demand analysis and maintenance in the project language. Not for: tasks that already require a fixed, complete analysis framework. This is a selectable behavior preference, not a universally best practice."
12
+ },
13
+ "scopes": {
14
+ "zh": [
15
+ "需要简明沟通、按需分析和项目语言维护的任务。"
16
+ ],
17
+ "en": [
18
+ "Tasks needing concise communication, on-demand analysis and maintenance in the project language."
19
+ ]
20
+ },
21
+ "rules": [
22
+ "om.conclusion-first@1.1.0",
23
+ "om.chinese-response@1.1.0",
24
+ "om.bilingual-terms@1.1.0",
25
+ "om.markdown-structure@1.1.0",
26
+ "om.clean-deliverable@1.1.0",
27
+ "om.concrete-titles@1.1.0",
28
+ "om.response-brief@1.1.0",
29
+ "om.evidence-inference@1.1.0",
30
+ "om.test-evidence@1.1.0",
31
+ "om.calculations@1.1.0",
32
+ "om.complete-options@1.1.0",
33
+ "om.align-requirements@1.1.0",
34
+ "om.batch-questions@1.1.0",
35
+ "om.assumptions-authority@1.1.0",
36
+ "om.solutions-proportional@1.1.0",
37
+ "om.versioned-files@1.1.0",
38
+ "om.device-readable@1.1.0",
39
+ "om.semantic-wrap@1.1.0",
40
+ "om.editable-diagrams@1.1.0",
41
+ "om.consistent-items@1.1.0",
42
+ "om.report-context@1.1.0",
43
+ "om.report-formats@1.1.0",
44
+ "om.report-review@1.1.0",
45
+ "om.tables-compact@1.1.0",
46
+ "om.portable-files@1.1.0",
47
+ "om.rules-tools@1.1.0",
48
+ "om.maintenance-project-language@1.1.0"
49
+ ],
50
+ "loading": {
51
+ "always": [
52
+ "om.chinese-response",
53
+ "om.conclusion-first",
54
+ "om.response-brief",
55
+ "om.evidence-inference",
56
+ "om.bilingual-terms",
57
+ "om.markdown-structure",
58
+ "om.batch-questions",
59
+ "om.align-requirements",
60
+ "om.clean-deliverable"
61
+ ],
62
+ "tags": [
63
+ {
64
+ "id": "decisions",
65
+ "name": {
66
+ "zh": "方案与论证",
67
+ "en": "Options and reasoning"
68
+ },
69
+ "trigger": {
70
+ "zh": "给方案、做比较、算数字、转述实测或依赖假设时",
71
+ "en": "Proposing or comparing options, computing numbers, relaying test results, or relying on assumptions"
72
+ },
73
+ "exclude": {
74
+ "zh": "纯执行类回复",
75
+ "en": "Pure execution"
76
+ }
77
+ },
78
+ {
79
+ "id": "files",
80
+ "name": {
81
+ "zh": "文件与条目",
82
+ "en": "Files and items"
83
+ },
84
+ "trigger": {
85
+ "zh": "改文件、命名版本、拟标题、做表格对比或整理文档时",
86
+ "en": "Editing or versioning files, writing titles, comparison tables, or organizing documents"
87
+ },
88
+ "exclude": {
89
+ "zh": "对话回复",
90
+ "en": "Chat replies"
91
+ }
92
+ },
93
+ {
94
+ "id": "reports",
95
+ "name": {
96
+ "zh": "汇报材料",
97
+ "en": "Report materials"
98
+ },
99
+ "trigger": {
100
+ "zh": "制作或审查汇报、PPT、HTML,含图表与排版时",
101
+ "en": "Producing or reviewing reports, decks or HTML with charts and layout"
102
+ },
103
+ "exclude": {
104
+ "zh": "闲聊与纯代码任务",
105
+ "en": "Chat and pure coding tasks"
106
+ }
107
+ },
108
+ {
109
+ "id": "maintenance",
110
+ "name": {
111
+ "zh": "规则与模板维护",
112
+ "en": "Rule and template upkeep"
113
+ },
114
+ "trigger": {
115
+ "zh": "编写或维护 Skill、规则、模板时",
116
+ "en": "Writing or maintaining skills, rules or templates"
117
+ },
118
+ "exclude": {
119
+ "zh": "代码与命令标识符",
120
+ "en": "Code and command identifiers"
121
+ }
122
+ }
123
+ ],
124
+ "members": {
125
+ "om.calculations": [
126
+ "decisions"
127
+ ],
128
+ "om.complete-options": [
129
+ "decisions"
130
+ ],
131
+ "om.solutions-proportional": [
132
+ "decisions"
133
+ ],
134
+ "om.assumptions-authority": [
135
+ "decisions"
136
+ ],
137
+ "om.test-evidence": [
138
+ "decisions"
139
+ ],
140
+ "om.versioned-files": [
141
+ "files"
142
+ ],
143
+ "om.concrete-titles": [
144
+ "files"
145
+ ],
146
+ "om.portable-files": [
147
+ "files"
148
+ ],
149
+ "om.consistent-items": [
150
+ "files"
151
+ ],
152
+ "om.tables-compact": [
153
+ "files"
154
+ ],
155
+ "om.report-context": [
156
+ "reports"
157
+ ],
158
+ "om.report-formats": [
159
+ "reports"
160
+ ],
161
+ "om.report-review": [
162
+ "reports"
163
+ ],
164
+ "om.device-readable": [
165
+ "reports"
166
+ ],
167
+ "om.semantic-wrap": [
168
+ "reports"
169
+ ],
170
+ "om.editable-diagrams": [
171
+ "reports"
172
+ ],
173
+ "om.rules-tools": [
174
+ "maintenance"
175
+ ],
176
+ "om.maintenance-project-language": [
177
+ "maintenance"
178
+ ]
179
+ }
180
+ }
181
+ }
@@ -0,0 +1,179 @@
1
+ {
2
+ "family": "scenario",
3
+ "variant": "reviewed-project",
4
+ "version": "1.1.0",
5
+ "name": {
6
+ "zh": "完整分析与项目交付组合",
7
+ "en": "Complete analysis with project delivery"
8
+ },
9
+ "description": {
10
+ "zh": "用途:需要完整分析、版本化交付及中文维护的项目任务。 不适用:未选择此偏好的项目;需要专属业务模板的任务。 这是可选择的行为偏好,不代表普遍最优做法。",
11
+ "en": "Use: project tasks needing complete analysis, versioned delivery and Chinese maintenance. Not for: projects that did not choose this preference; tasks needing business-specific templates. This is a selectable behavior preference, not a universally best practice."
12
+ },
13
+ "scopes": {
14
+ "zh": [
15
+ "需要完整分析、版本化交付及中文维护的项目任务。"
16
+ ],
17
+ "en": [
18
+ "Project tasks needing complete analysis, versioned delivery and Chinese maintenance."
19
+ ]
20
+ },
21
+ "rules": [
22
+ "om.conclusion-first@1.1.0",
23
+ "om.chinese-response@1.1.0",
24
+ "om.bilingual-terms@1.1.0",
25
+ "om.markdown-structure@1.1.0",
26
+ "om.clean-deliverable@1.1.0",
27
+ "om.concrete-titles@1.1.0",
28
+ "om.response-dense@1.1.0",
29
+ "om.evidence-inference@1.1.0",
30
+ "om.test-evidence@1.1.0",
31
+ "om.calculations@1.1.0",
32
+ "om.complete-options@1.1.0",
33
+ "om.align-requirements@1.1.0",
34
+ "om.batch-questions@1.1.0",
35
+ "om.assumptions-authority@1.1.0",
36
+ "om.solutions-three-level@1.1.0",
37
+ "om.versioned-files@1.1.0",
38
+ "om.device-readable@1.1.0",
39
+ "om.semantic-wrap@1.1.0",
40
+ "om.editable-diagrams@1.1.0",
41
+ "om.consistent-items@1.1.0",
42
+ "om.report-context@1.1.0",
43
+ "om.report-formats@1.1.0",
44
+ "om.report-review@1.1.0",
45
+ "om.tables-list-first@1.1.0",
46
+ "om.portable-files@1.1.0",
47
+ "om.rules-tools@1.1.0",
48
+ "om.maintenance-chinese@1.1.0"
49
+ ],
50
+ "loading": {
51
+ "always": [
52
+ "om.chinese-response",
53
+ "om.conclusion-first",
54
+ "om.response-dense",
55
+ "om.evidence-inference",
56
+ "om.bilingual-terms",
57
+ "om.markdown-structure",
58
+ "om.batch-questions",
59
+ "om.align-requirements",
60
+ "om.clean-deliverable",
61
+ "om.tables-list-first"
62
+ ],
63
+ "tags": [
64
+ {
65
+ "id": "decisions",
66
+ "name": {
67
+ "zh": "方案与论证",
68
+ "en": "Options and reasoning"
69
+ },
70
+ "trigger": {
71
+ "zh": "给方案、做比较、算数字、转述实测或依赖假设时",
72
+ "en": "Proposing or comparing options, computing numbers, relaying test results, or relying on assumptions"
73
+ },
74
+ "exclude": {
75
+ "zh": "纯执行类回复",
76
+ "en": "Pure execution"
77
+ }
78
+ },
79
+ {
80
+ "id": "files",
81
+ "name": {
82
+ "zh": "文件与条目",
83
+ "en": "Files and items"
84
+ },
85
+ "trigger": {
86
+ "zh": "改文件、命名版本、拟标题或整理文档时",
87
+ "en": "Editing or versioning files, writing titles, or organizing documents"
88
+ },
89
+ "exclude": {
90
+ "zh": "对话回复",
91
+ "en": "Chat replies"
92
+ }
93
+ },
94
+ {
95
+ "id": "reports",
96
+ "name": {
97
+ "zh": "汇报材料",
98
+ "en": "Report materials"
99
+ },
100
+ "trigger": {
101
+ "zh": "制作或审查汇报、PPT、HTML,含图表与排版时",
102
+ "en": "Producing or reviewing reports, decks or HTML with charts and layout"
103
+ },
104
+ "exclude": {
105
+ "zh": "闲聊与纯代码任务",
106
+ "en": "Chat and pure coding tasks"
107
+ }
108
+ },
109
+ {
110
+ "id": "maintenance",
111
+ "name": {
112
+ "zh": "规则与模板维护",
113
+ "en": "Rule and template upkeep"
114
+ },
115
+ "trigger": {
116
+ "zh": "编写或维护 Skill、规则、模板时",
117
+ "en": "Writing or maintaining skills, rules or templates"
118
+ },
119
+ "exclude": {
120
+ "zh": "代码与命令标识符",
121
+ "en": "Code and command identifiers"
122
+ }
123
+ }
124
+ ],
125
+ "members": {
126
+ "om.calculations": [
127
+ "decisions"
128
+ ],
129
+ "om.complete-options": [
130
+ "decisions"
131
+ ],
132
+ "om.solutions-three-level": [
133
+ "decisions"
134
+ ],
135
+ "om.assumptions-authority": [
136
+ "decisions"
137
+ ],
138
+ "om.test-evidence": [
139
+ "decisions"
140
+ ],
141
+ "om.versioned-files": [
142
+ "files"
143
+ ],
144
+ "om.concrete-titles": [
145
+ "files"
146
+ ],
147
+ "om.portable-files": [
148
+ "files"
149
+ ],
150
+ "om.consistent-items": [
151
+ "files"
152
+ ],
153
+ "om.report-context": [
154
+ "reports"
155
+ ],
156
+ "om.report-formats": [
157
+ "reports"
158
+ ],
159
+ "om.report-review": [
160
+ "reports"
161
+ ],
162
+ "om.device-readable": [
163
+ "reports"
164
+ ],
165
+ "om.semantic-wrap": [
166
+ "reports"
167
+ ],
168
+ "om.editable-diagrams": [
169
+ "reports"
170
+ ],
171
+ "om.rules-tools": [
172
+ "maintenance"
173
+ ],
174
+ "om.maintenance-chinese": [
175
+ "maintenance"
176
+ ]
177
+ }
178
+ }
179
+ }
@@ -0,0 +1,31 @@
1
+ {
2
+ "family": "work-method",
3
+ "variant": "proportional",
4
+ "version": "1.1.0",
5
+ "name": {
6
+ "zh": "证据核验与按需分析",
7
+ "en": "Evidence checks with proportional analysis"
8
+ },
9
+ "description": {
10
+ "zh": "用途:按影响和复杂度选择分析深度的问题解决。 不适用:用户已要求固定层次或完整框架的任务。 这是可选择的行为偏好,不代表普遍最优做法。",
11
+ "en": "Use: problem solving where analysis depth follows impact and complexity. Not for: tasks where the user already asked for fixed levels or a complete framework. This is a selectable behavior preference, not a universally best practice."
12
+ },
13
+ "scopes": {
14
+ "zh": [
15
+ "按影响和复杂度选择分析深度的问题解决。"
16
+ ],
17
+ "en": [
18
+ "Problem solving where analysis depth follows impact and complexity."
19
+ ]
20
+ },
21
+ "rules": [
22
+ "om.evidence-inference@1.1.0",
23
+ "om.test-evidence@1.1.0",
24
+ "om.calculations@1.1.0",
25
+ "om.complete-options@1.1.0",
26
+ "om.align-requirements@1.1.0",
27
+ "om.batch-questions@1.1.0",
28
+ "om.assumptions-authority@1.1.0",
29
+ "om.solutions-proportional@1.1.0"
30
+ ]
31
+ }
@@ -0,0 +1,31 @@
1
+ {
2
+ "family": "work-method",
3
+ "variant": "three-level",
4
+ "version": "1.1.0",
5
+ "name": {
6
+ "zh": "证据核验与三层方案",
7
+ "en": "Evidence checks with three-level solutions"
8
+ },
9
+ "description": {
10
+ "zh": "用途:分析、比较与问题解决。 不适用:纯执行且无需方案分析的任务;用户明确排除的分支。 这是可选择的行为偏好,不代表普遍最优做法。",
11
+ "en": "Use: analysis, comparison and problem solving. Not for: pure execution with no solution analysis; branches the user explicitly excluded. This is a selectable behavior preference, not a universally best practice."
12
+ },
13
+ "scopes": {
14
+ "zh": [
15
+ "分析、比较与问题解决。"
16
+ ],
17
+ "en": [
18
+ "Analysis, comparison and problem solving."
19
+ ]
20
+ },
21
+ "rules": [
22
+ "om.evidence-inference@1.1.0",
23
+ "om.test-evidence@1.1.0",
24
+ "om.calculations@1.1.0",
25
+ "om.complete-options@1.1.0",
26
+ "om.align-requirements@1.1.0",
27
+ "om.batch-questions@1.1.0",
28
+ "om.assumptions-authority@1.1.0",
29
+ "om.solutions-three-level@1.1.0"
30
+ ]
31
+ }
@@ -3,7 +3,7 @@
3
3
  "packs": [
4
4
  {
5
5
  "id": "opinion-manager/catalog",
6
- "version": "1.0.0",
6
+ "version": "1.1.0",
7
7
  "kind": "catalog",
8
8
  "title": "公开 Opinion 规则与模板目录",
9
9
  "description": "可选择的规则与模板版本;项目采用情况由 opinion.lock.json 记录,更新经 opinion-manager 预览与确认。",
@@ -13,50 +13,91 @@
13
13
  "materializes": [],
14
14
  "project_lock": "opinion-manager",
15
15
  "files": {
16
- "catalog/README.md": "af654d84bdbc22c020dda7fe090d0ee364d5a892e93bdfa5fdd6c3305635ccb7",
16
+ "catalog/README.md": "246b5c9896eb0b8a1187626ecabb113aec48de0060e6f8f3612cc1a8e04d970c",
17
17
  "catalog/rules/.gitkeep": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
18
18
  "catalog/rules/om.align-requirements/versions/1.0.0.json": "26b95f2edce0d4942b3652c624b270a29fd75caa1fc5717debdffca126241091",
19
+ "catalog/rules/om.align-requirements/versions/1.1.0.json": "13b83b21cf168b7aac9324f119781f8a7ef167817fccfc344248c06ad0b55042",
19
20
  "catalog/rules/om.assumptions-authority/versions/1.0.0.json": "85b02b03bbf8447f703820026d9ebe6e9ac1182000c90451dad4c81cd23c23d6",
21
+ "catalog/rules/om.assumptions-authority/versions/1.1.0.json": "9b09876d87f31e96e6fc72327b00f42ef33a75ced6a90aa8056c1c4a853891a5",
20
22
  "catalog/rules/om.batch-questions/versions/1.0.0.json": "2cad060df9195bf6f47083b0443fb92a9e1735989991aab34aae0c2651d18337",
23
+ "catalog/rules/om.batch-questions/versions/1.1.0.json": "3a94132fb0b3c2ddde4d475460eb27eee07de016474ecdb71ee79f80d80720d2",
21
24
  "catalog/rules/om.bilingual-terms/versions/1.0.0.json": "420531760d1ab5ffe52395a680597cb652574027e542fd4684b7166e14918c39",
25
+ "catalog/rules/om.bilingual-terms/versions/1.1.0.json": "a4088be89abc358e845ede0059d48e561e88db32244f3ff61e5f95315137712f",
22
26
  "catalog/rules/om.calculations/versions/1.0.0.json": "e54f87a7993d5908a1adfff07534b85af96a317d90b6eda6b84d3f8701331a18",
27
+ "catalog/rules/om.calculations/versions/1.1.0.json": "a63091197230f82018cb0d806cf316f047f56e45f3a60e4ddfea4704792d9244",
23
28
  "catalog/rules/om.chinese-response/versions/1.0.0.json": "6bd3337a65a81251fad64555a7702016807c64e78aeb9709cb50cdc7325662c5",
29
+ "catalog/rules/om.chinese-response/versions/1.1.0.json": "096dc9a1446ef259349f0cd26b284cfa62a3ce3dab445e7a0bcd44cdb0386c65",
24
30
  "catalog/rules/om.clean-deliverable/versions/1.0.0.json": "f0af00b2666ca7146447f17c8e13ca01c9e0d73c01883797787b44c045b06df2",
31
+ "catalog/rules/om.clean-deliverable/versions/1.1.0.json": "8d75b649d05a5c774ccde496708ec52fa369e1843d7138985585e8d9f04ec033",
25
32
  "catalog/rules/om.complete-options/versions/1.0.0.json": "ef9ee103ee35dea554ae5d72735ebdeb7b0e16bf707b2362ef31be004c63110b",
33
+ "catalog/rules/om.complete-options/versions/1.1.0.json": "61f68e08a26a7f05df0d93b784135d5f45883c025b14049e50ca076c6c4f3000",
26
34
  "catalog/rules/om.conclusion-first/versions/1.0.0.json": "2c0975c3fba8d4553b69a714fd324a1240c10ba0429904740db853cf10d8f8d3",
35
+ "catalog/rules/om.conclusion-first/versions/1.1.0.json": "56fb508b615bfcd87ccb3fe28e7a3ad8e3214ecf98b8a10d0c3e8ff1158963bc",
27
36
  "catalog/rules/om.concrete-titles/versions/1.0.0.json": "019d017fc9181339f1f969fd3e81323761e28b755c4ef434fb6af4d55925f538",
37
+ "catalog/rules/om.concrete-titles/versions/1.1.0.json": "0b73490415ee802a7f56210f95fb4a16aadf317b3d7b0a1b454c8668d4b58c3a",
28
38
  "catalog/rules/om.consistent-items/versions/1.0.0.json": "dd7030332d69d2eb91a7b1af34a3ac37d49e4018365da1e7a2d10afd71ed66f0",
39
+ "catalog/rules/om.consistent-items/versions/1.1.0.json": "469d766d152c5e6016b40f819210def87431513d02c6af156dcf547c146e7070",
29
40
  "catalog/rules/om.device-readable/versions/1.0.0.json": "0afce0068b778ffa8c623cc1ff208192f70e03434f0ace4fd303565ba728a75b",
41
+ "catalog/rules/om.device-readable/versions/1.1.0.json": "4119293934f793261f9de6c0b284eaecfbbc631f097d962c5fef033f00acbabb",
30
42
  "catalog/rules/om.editable-diagrams/versions/1.0.0.json": "68d293ebd061527d948e892b9f4503a74043ed37ecf2a922cdd9047f508e2ed8",
43
+ "catalog/rules/om.editable-diagrams/versions/1.1.0.json": "d99260419700ca8470f2d2aea5833e8fd3250c1565ee8aad22c9491a271a3d29",
31
44
  "catalog/rules/om.evidence-inference/versions/1.0.0.json": "01593b17d08a67ee41b28d96e454b25d752fd5eccc341476a75d2bfb541884f2",
45
+ "catalog/rules/om.evidence-inference/versions/1.1.0.json": "f44065db407191f229375f1a76a282468618e5dbd39dd014c4bbc6a33856b60c",
32
46
  "catalog/rules/om.maintenance-chinese/versions/1.0.0.json": "6a6659917a0bccda3750f79727ea433eb63d40426aa2a212b65499a8160fc181",
47
+ "catalog/rules/om.maintenance-chinese/versions/1.1.0.json": "a2ba7d257ec73f3ff8aae12407aab79e0f76e2c58ac30fd2607b14b0939f0218",
33
48
  "catalog/rules/om.maintenance-project-language/versions/1.0.0.json": "c29d447ac315533a52a6c242d7cc52b1668d9945c87c1f922dce261872730235",
49
+ "catalog/rules/om.maintenance-project-language/versions/1.1.0.json": "ce1cbd2b5cf976dbb618ff2b0a03fd4a2c75ea7cd90fd9cf5b3f4e7ea97f4504",
34
50
  "catalog/rules/om.markdown-structure/versions/1.0.0.json": "3647daeb2f4e4f532c8f77a033703bd3b36bbdd355a1f1f6ce8ac1e19afb66d7",
51
+ "catalog/rules/om.markdown-structure/versions/1.1.0.json": "5e026f87da514f3bdb3c883bc1cd202e0414dc656f0021e35b47223b76f48a9a",
35
52
  "catalog/rules/om.portable-files/versions/1.0.0.json": "f638ce7554a11e8f793b63ee01ea0ca65a8df405b3ca34da1ad8865960908057",
53
+ "catalog/rules/om.portable-files/versions/1.1.0.json": "d2af355593b8fffba6d19b558bb1edf23bf367f647af2d938d3cb9b1565ef0f1",
36
54
  "catalog/rules/om.report-context/versions/1.0.0.json": "c423df33e1ad8dbf7d060b15fb23b35b68ebf924beb76ac129b346e5cf46c917",
55
+ "catalog/rules/om.report-context/versions/1.1.0.json": "4c6f467b8fe75f878d4f946f8a8f88049d3a28dddbc18368c5705f84e4b47cb2",
37
56
  "catalog/rules/om.report-formats/versions/1.0.0.json": "1fa5a40415061366227cdba554bc811bfabdc764b110adfdfedf5935e2239774",
57
+ "catalog/rules/om.report-formats/versions/1.1.0.json": "9ad59f85e307b0db19bdfb0d8bc31e26d39fb70523ea246980e7ab00a70f1e09",
38
58
  "catalog/rules/om.report-review/versions/1.0.0.json": "054ebce7547cf7254a0d9d03af8b651daf6ae6361c921e3fa111aea7f9d96d4e",
59
+ "catalog/rules/om.report-review/versions/1.1.0.json": "c607bd8730f9a48fa7dd0be4d7c404454618eb8f0340b698c3c3acec0784cb91",
39
60
  "catalog/rules/om.response-brief/versions/1.0.0.json": "1d64ac6bedde9fd7205c6e517bb484acebedd08fe82f06aa761cf04bf010414a",
61
+ "catalog/rules/om.response-brief/versions/1.1.0.json": "f50c23c0f658e9a2d14f491413e3f2819fba1eb4d8e5ba6610c166bc89b6da22",
40
62
  "catalog/rules/om.response-dense/versions/1.0.0.json": "5dda9f066dae841df986a190f00ea0bc56c0d3b0ce42ebffa9b2e1910dc0c4b9",
63
+ "catalog/rules/om.response-dense/versions/1.1.0.json": "d77749efa3632c92ff4578b2c5fac0ec4f0e5d11649898a0e4a81908e212be26",
41
64
  "catalog/rules/om.rules-tools/versions/1.0.0.json": "6f0657b70bd2e15e0dc68b95a9ef77234cd4af79d40a9a9f8a8158da16323a8d",
65
+ "catalog/rules/om.rules-tools/versions/1.1.0.json": "230320a6db7fa104d10397835144980b2bb6f2d9764e1f61896e4a4fecd71575",
42
66
  "catalog/rules/om.semantic-wrap/versions/1.0.0.json": "77890d37d6e21bf0568df250d5ec6398699c595728decf624d2f6ac8887802d7",
67
+ "catalog/rules/om.semantic-wrap/versions/1.1.0.json": "109922c31ce787651ce8bdacb417d2a55eb59207b82a665a7867634bfbb899a6",
43
68
  "catalog/rules/om.solutions-proportional/versions/1.0.0.json": "4ba50387659bc4e4593026a2ab253abda0f9d0e8e959063957dc58d72dbd5dac",
69
+ "catalog/rules/om.solutions-proportional/versions/1.1.0.json": "270405661e4e8bc4e553da72cbd0b101c2b2dad83deb121e657f2f64612f81c3",
44
70
  "catalog/rules/om.solutions-three-level/versions/1.0.0.json": "c9c2c2bad29e79a649ce04c95c9a2cb5a7ba6b92e5318878d7d5a40e50df7b6d",
71
+ "catalog/rules/om.solutions-three-level/versions/1.1.0.json": "e092af0ea720db5d59d727e778a01f1251aa8dd69a3f1cb1d4f3befd783b2838",
45
72
  "catalog/rules/om.tables-compact/versions/1.0.0.json": "0822de141431d920e6bf8ae0cc7a07641ad34e4248d252133f5a488fc17610f4",
73
+ "catalog/rules/om.tables-compact/versions/1.1.0.json": "d292b63d74d8dcb83516b9031afe4f60402b2e4e4c1db8bc2f9275431988dd02",
46
74
  "catalog/rules/om.tables-list-first/versions/1.0.0.json": "8e4d9f8d0382ca9a4a06a31e64bc402bfe1d2f8fd80f576670d28de26565fe15",
75
+ "catalog/rules/om.tables-list-first/versions/1.1.0.json": "27a37e8d5265f16aabed620bb4b6084707b5bd8412f3c7d55ad420ffd6cec1da",
47
76
  "catalog/rules/om.test-evidence/versions/1.0.0.json": "5fb88ea8478601128a5662077beb8180126c3c7529a9e54fa4b324a47e1b9304",
77
+ "catalog/rules/om.test-evidence/versions/1.1.0.json": "e326dc4f447ff22ee270b7271bb9dc3414888b4e7f2c5ae3a889be58102ac630",
48
78
  "catalog/rules/om.versioned-files/versions/1.0.0.json": "b5cc996847cf20f4764be135c4a5189a30f16dd26d4eaeda38bf6318820c256c",
79
+ "catalog/rules/om.versioned-files/versions/1.1.0.json": "3e8cbf610d5d5e1db63bf5c8fc5b32adf1d0489216cfc3248146a688fe17bb2d",
49
80
  "catalog/templates/.gitkeep": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
50
81
  "catalog/templates/delivery-standard/compact-tables/versions/1.0.0.json": "b0d988b5e62017ce4784466f590fdc3ac3693412541ab71f788beac0658c25e1",
82
+ "catalog/templates/delivery-standard/compact-tables/versions/1.1.0.json": "f0b724526a56c6aa9f22388a2696c093ae6d33fb7c04325413818def2008dd03",
51
83
  "catalog/templates/delivery-standard/list-first/versions/1.0.0.json": "3ea2774fb2f3fb5943d83eae3d36c637b4bcc6bbc45c54cf4428de0db1c4240f",
84
+ "catalog/templates/delivery-standard/list-first/versions/1.1.0.json": "f998047856cf540ea66ef67146c10d4562f16d0e97b90d00184538fe97548034",
52
85
  "catalog/templates/development-standard/chinese-maintenance/versions/1.0.0.json": "3d268a6a323d4bac1afcd5b8539b65f4019ab112850b223c61ee06615840248d",
86
+ "catalog/templates/development-standard/chinese-maintenance/versions/1.1.0.json": "8e2cb128b93efc3e6c04c418fc68a16c32eb200b7231d3829b3a4c3423cd17d3",
53
87
  "catalog/templates/development-standard/project-language/versions/1.0.0.json": "6f2d69873b198ce0cf7fac509cbb8f4702de37c8d98ff3c5ec31578022e0d990",
88
+ "catalog/templates/development-standard/project-language/versions/1.1.0.json": "5f0e8750efdabbf4b3de84258c1f978d103d60afcc157798b4271468595239fe",
54
89
  "catalog/templates/response-style/brief-chinese/versions/1.0.0.json": "951aacd484989835be5811730510a103d2efe044342f275066f64765d747448b",
90
+ "catalog/templates/response-style/brief-chinese/versions/1.1.0.json": "6e1fd5bf2d343ae3f769ba5df60e2a9754a147935a3fbd12272ff2488d956264",
55
91
  "catalog/templates/response-style/dense-chinese/versions/1.0.0.json": "72479e811a3db0a9c50401a60f0c06694e00014a558ad1c20ed78945e99edcfa",
92
+ "catalog/templates/response-style/dense-chinese/versions/1.1.0.json": "8c2a06377aba9876ab875c32e9df33cf508dc2b5ac05b10d3363c68c07c36b26",
56
93
  "catalog/templates/scenario/concise-project/versions/1.0.0.json": "efa1adfed739a206699cd6ae34b620d8c4717343959ffe63fe7ecc0456b16614",
94
+ "catalog/templates/scenario/concise-project/versions/1.1.0.json": "6ead723134faf73b2c63ea5f73b89b460bdade4a6c850ad50459450da152fcea",
57
95
  "catalog/templates/scenario/reviewed-project/versions/1.0.0.json": "4b51fcccb57c237eee3fff56b2a73c86252bb7ee9cb5e71d1aa2cfc0d6062aa1",
96
+ "catalog/templates/scenario/reviewed-project/versions/1.1.0.json": "056c2da8c8bcacfce3fe4da1c41e4c0c1715b1c8a5121a102e23b0b772359b26",
58
97
  "catalog/templates/work-method/proportional/versions/1.0.0.json": "e73ea29f532d09813b00ed161e9beb7ab079eb56dda5ecffda1c30cc630d0bd8",
59
- "catalog/templates/work-method/three-level/versions/1.0.0.json": "0fa27605f23901f7d120d5d6c4ee63f14dbe12c4da53b1e3287da3d4b68da5ef"
98
+ "catalog/templates/work-method/proportional/versions/1.1.0.json": "12724039f6f7a45d9846549a638c6db78daba3b156fd0bde5d4704687c951cf1",
99
+ "catalog/templates/work-method/three-level/versions/1.0.0.json": "0fa27605f23901f7d120d5d6c4ee63f14dbe12c4da53b1e3287da3d4b68da5ef",
100
+ "catalog/templates/work-method/three-level/versions/1.1.0.json": "1d594f1d70693d5c2a7498293c08b81d32cc23fa8609da3f163802f57f60976f"
60
101
  }
61
102
  }
62
103
  ]
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
3
3
  "name": "opinion-manager",
4
- "version": "0.6.0",
4
+ "version": "0.7.0",
5
5
  "description": "Manage Agent behavior rules through templates, partial selection, blank iteration, and authorized-memory candidates, with complete previews and exact version locks.",
6
6
  "author": {
7
7
  "name": "lc4t"
@@ -23,19 +23,40 @@ The inbox keeps opinions that the Agent noticed during work and that still await
23
23
 
24
24
  All commands are `opinion_manager.py inbox <action>`, and `--project` is required. `--lang zh|en` selects the message language (by default `AGENTHELM_LANG`, then the system locale, then English). Output is JSON; `list --output markdown` gives a numbered list for the user.
25
25
 
26
- - `add --title --text --scope [--quote] [--client] [--session]`: stash a candidate. Title up to 120, text up to 1200 and quote up to 600 characters; at least one `--scope`. Candidates with the same content (ignoring whitespace and case) are merged and the new quote is added as evidence.
27
- - `list [--status a,b] [--session ID] [--unpresented]`: list candidates.
26
+ - `add --title --text --scope --trigger --purpose [--exclude] [--relation] [--quote] [--client] [--session]`: stash a candidate. Title up to 120, text up to 1200, quote up to 600 and each other field up to 300 characters; at least one `--scope`. Candidates with the same content (ignoring whitespace and case) are merged and the new quote is added as evidence.
27
+ - `--trigger` (when it applies) and `--purpose` (why) are filled in by the agent at stash time, so the user sees them once when deciding in a batch. In this version they are **soft-required**: when missing the stash still succeeds, returns `warnings` and `complete: false`, and listings show the candidate as incomplete. Rejecting missing fields is left to a later major (one switch in code: `STRICT_REQUIRED`).
28
+ - A value that is passed but invalid is always rejected: blank, or a placeholder such as "all", "everything", "relevant", "n/a" (and the Chinese equivalents).
29
+ - Global declaration: when the user explicitly asks for a global rule, `--trigger` must be the fixed standard value "All scenarios; the user asked for this to apply globally" (Chinese: 「全部场景,用户要求全局」) together with a non-empty `--quote`; listings mark it "Global (ask the user to confirm)". **The tool only checks the form and cannot verify who asked for it to be global**; the user confirms it when deciding.
30
+ - `--exclude` states when it does not apply; `--relation` is the agent's own note on how this relates to existing rules, stored and shown verbatim, never interpreted.
31
+ - When adopted, the custom line is `- **Title** (Applies to: ...; When: ...; Purpose: ...; Not for: ...): text` (older candidates without these fields render byte-for-byte as before). The labels on the line and the frame text of `OPINION.md` follow the **project file language** (`language` in the Project Profile, default zh), not the interface language; an `en` project file contains no Chinese except what the user wrote. For a project whose `language` is not zh, `opinion.lock.json` carries an optional `"language": "en"` so `verify` can reproduce the rendering; the lock format of zh projects is unchanged. Older runtimes cannot read a lock that carries `language`; upgrade first.
32
+ - `list [--status a,b] [--session ID] [--unpresented]`: list candidates. Each JSON item carries `relations` (mechanical relation hints; an empty list when none), and markdown shows them as "Possibly related" lines.
33
+ - Hints compare against: other candidates (rejected ones excluded), the custom lines of `OPINION.md` (parsed in this tool's own two render formats), and the public rule texts selected in the Profile. Three kinds: equal normalized text (`same_text`), character-bigram Jaccard overlap >= 0.5 (`text_overlap`; both sides at least 6 characters after normalization), trigger overlap >= 0.4 (`condition_overlap`; global declarations excluded). Each carries an evidence snippet and "Mechanical comparison, for reference only", at most 5, never blocking; thresholds are conservative and prefer missing a hint. Semantic judgments (conflict, merge, widen/narrow, split) belong to the agent, who writes them into the `--relation` note; the tool stores and shows it verbatim and never interprets it.
34
+ - `check`: read-only, writes no file. Lists candidates (`pending`, `trial`) missing trigger / purpose and the custom lines of `OPINION.md`; lines in an unrecognized format are marked `unstructured` and only reported. Each candidate comes with a `revise_command` so the agent can propose values one by one. Completing lines in `OPINION.md` goes through the full `adopt` preview flow, confirmed by the user in one batch; there is no new write command.
35
+ - `revise <id> [--title --text --scope --trigger --purpose --exclude --relation --quote]`: revise a candidate. When the status is `pending` and it was never shown (`presented` is 0) no `--confirmed` is needed; in every other status `--confirmed` is required. `history` gets `{to: the unchanged status, note: revised, previous: old values of the changed fields}`; the status stays; the digest is recomputed and a collision with another candidate is an error; `--scope` replaces the list, `--quote` appends. Revising a `formal` candidate does not change `OPINION.md` (a warning is returned).
28
36
  - `decide <ids...> --to pending|trial|formal|held|rejected --confirmed [--note]`: record a decision. Without `--confirmed` it is always refused; with several ids, either all succeed or none changes.
29
37
  - `adopt <ids...> [--apply --confirm SHA256]`: the one-confirmation path from "adopt" to the written rule. Use it when the user has said "adopt" in the conversation; no separate `decide`.
30
38
  - Without `--apply`: preview only, nothing is written. The candidate text is appended as personal rules to the custom section of the project's existing Profile (a new `private.my-opinion@1.0.0` if there is none) and goes through the same preview as `profile` and `compose`. It returns the full `content` (what `OPINION.md` will contain), `blocked` (for example hand-written content that must be handled first) and `confirmation_sha256`.
39
+ - The preview also returns `added_lines` (the lines that will be appended to the custom section, identical to those in `content`) and `shown_matches`: `true` when every candidate has a shown snapshot, each snapshot equals the line rendered now, nothing is `blocked` and there is no conflict; `shown` lists the result per candidate. It is a record of fact and does not relax the protection: writing still needs the preview fingerprint, and `shown_matches` loosens nothing.
31
40
  - With `--apply --confirm <the preview's confirmation_sha256>`: once the user has confirmed, it writes a new Profile version, `OPINION.md` and `opinion.lock.json` in one go, reads them back with `verify`, and only then marks the candidates `formal`. If candidates, the Profile or `OPINION.md` changed since the preview, the fingerprint no longer matches and nothing is written. It does not use `--confirmed`: the confirmation is the preview fingerprint.
32
41
  - Several ids share one preview and one confirmation, and either all succeed or none changes; a missing id or an id that is already `formal` fails the whole call and is named. If writing or the read-back fails, `OPINION.md`, the lock and the Profile file created by this run are restored and the candidate status is unchanged.
33
42
  - A candidate that `decide --to formal` already recorded but that is not written yet: move it back with `decide --to pending --confirmed`, then `adopt`.
34
- - `present <ids...>`: record that the user has been shown these candidates (the hook calls it when it sends the confirmation request, so the same candidate is not raised again and again).
43
+ - `present <ids...>`: record that the user has been shown these candidates (the hook calls it when it sends the confirmation request, so the same candidate is not raised again and again). It also records a **shown snapshot**: the digest of the rule line the candidate would be rendered as at that moment (`shown`: `at`, `language`, `line_sha256`). If the candidate is revised or the project language changes afterwards, the digest no longer matches. The output carries these candidates (with `relations`).
44
+ - `gate [--session ID] [--cooldown-hours N] [--max-per-session N] [--peek]`: ask throttling; it only does the bookkeeping and does not decide whether to ask. At most one ask per project within N hours (default 24) and at most `--max-per-session` asks per session (default 1); returns `allowed`, `reason` (`ok` / `cooldown` / `session_limit`) and `next_allowed_at`. An allowed ask is recorded only without `--peek`; `--peek` is read-only and creates no file. Without `--session` only the cooldown applies. State lives in the project-private `.opinion/ask-state.json` (the last 50 sessions are kept); symbolic links are refused and read-modify-write holds `.opinion/.ask.lock`; a malformed state file is an error, not an allow. The command only keeps counts; when to call it and how to word the question is up to the protocol and hook.
35
45
  - `expire --days N`: return trials older than N days to `pending`. The return keeps `last_action_at`, so the item still counts as "untouched for a long time".
36
46
  - `pending [--session ID]`: the pending-item provider, see below.
37
47
  - `exposure`: read-only; reports the Git state of `.opinion/inbox.json` and the reminder text.
38
- - `brief [--session ID] [--max-chars N]`: builds the protocol and the full text of trial candidates injected at session start; the default limit is 6000 characters.
48
+ - `brief [--session ID] [--max-chars N] [--mode ask|digest|quiet] [--cooldown-hours N] [--max-per-session N]`: builds the protocol and the full trial candidates injected at session start (default limit 6000 characters). **This protocol is the single source for when to confirm** (SKILL.md only points to it). `--mode`: `ask` (default) tells the agent to run `gate` before asking at the end of a task, ask only when allowed, and `present` while showing; `digest` never asks separately and adds one closing sentence to a reply; `quiet` relies on the session-start summary alone. `--cooldown-hours` defaults to 24 and `--max-per-session` to 1; each goes into the `gate` command line of the `ask` protocol only when it differs from the default (the same as gate's own default, so omitting it is equivalent). In all three modes the protocol says: when the `adopt` preview has `shown_matches` true, no `blocked` and no conflicts, the agent may treat it as the user having confirmed what was shown and run `--apply --confirm` directly (the fingerprint protection is unchanged), then tell the user the result and how to go back (the previous Profile version is still under `.opinion/profiles/`; there is no one-step undo command); otherwise it shows the preview differences and asks once more. The user later updating a requirement does not itself trigger a confirmation.
49
+ - **Limits**: the hook path (agent-pack's `nudge`) has hard limits: `stop` fires at most once per session and asks `gate` first. The model-judged path (clients without hooks, or the model deciding when to ask) has no hard limit and relies on the protocol wording alone; this tool does not guarantee it. Real client behavior (Cursor sessionStart injection, Codex `/hooks` trust, Claude Stop `additionalContext` across versions) has not been verified on real clients.
50
+
51
+ ## Route rules (send a kind of task to a skill)
52
+
53
+ A rule can say "when doing this kind of work, use that skill", for example "all text deliverables finally go through a de-AI-flavor skill". **A route is only text in a rule: it raises the chance that the model calls the skill but cannot guarantee it**; never describe it to the user as "guaranteed" or "enforced". It adds to the skill's own `description` trigger rather than replacing it.
54
+
55
+ - Route options of `add` / `revise` (all optional): `--route-skill NAME` (a skill name: lowercase letters, digits, hyphens, at most 64; `plugin:skill` is not accepted), `--route-when before|after|instead` (default `before`), `--route-final` (the closing step: only with `after`, and it implies `after` when `--route-when` is not given), `--route-on TAG` (a short tag: lowercase letters, digits, hyphens, at most 32, used for mechanical comparison). The other route options need `--route-skill`; `revise` merges into the candidate's existing route, and `--route-clear` removes the whole route (exclusive with the other `--route-*` options; the old route goes into `previous` in `history`; confirmation rules are those of any revision). Conditions, purpose and exclusions still use `--trigger` / `--purpose` / `--exclude`; for "all text work" write the trigger as "producing a text deliverable for people" and list code, configuration, command output, keep-verbatim and so on under `--exclude`; do not make it global.
56
+ - The line written on adopt ends with a pure-ASCII marker `[route skill=… when=… final=1 on=…]` (`final` and `on` are omitted when unset), independent of the project language; the renderer adds two constraints in the **project file language**: the exclusion gains "the execution of the skill itself", and the text gains "while that skill runs, this rule does not trigger again" and "if that skill is unavailable, tell the user and complete the work yourself according to the purpose". Rule lines without a route are byte-identical to before; a `language=en` project has no Chinese. The Profile schema and the lock format do not change.
57
+ - A candidate's `route` object holds `skill`, `when`, `final`, `on`; older records lack the key. `list`, `present`, `pending`, the trial entries of `brief` and the `adopt` preview all show the route; the adopt fingerprint and the shown snapshot change together with the rendered line.
58
+ - `routes list [--output markdown]`: read-only. Lists candidates (`pending`, `trial`, and approved `formal` ones not yet written) and the routes in the custom section of `OPINION.md`; lines it does not recognize are marked `unstructured` and only reported. It adds **mechanical checks, for reference only** (warnings, never blocking): the same `--route-on` tag and timing routed to different skills; duplicate routes (same skill, tag and timing); more than one `--route-final` in the project; more than 6 routes; a single rule over 500 bytes. It **does not check whether a skill exists** (that needs `agenthelm inventory`, planned for the next release) and does not judge semantic overlap or indirect cycles (the `unchecked` field lists them honestly).
59
+ - Verification (this version): when submitting the lifecycle `use` event at the end of a task, self-report each applicable route rule in the existing `results` field as `passed` (the skill was actually called and used), `not-applicable` or `violated` (it should have been called but was not). Self-reporting depends on the agent recording honestly, and no route-adherence threshold has been validated; lifecycle is unchanged.
39
60
 
40
61
  ## Pending-item provider contract
41
62