skillpack-one 0.1.0-alpha.10

package/skills/atom-design-interface-spec/skill.contract.yaml

@@ -0,0 +1,26 @@
+schemaVersion: 1
+id: atom-design-interface-spec
+kind: atom
+version: 0.1.0
+name: { en: Design an interface specification, zh-CN: 设计界面规格 }
+summary: { en: Define an implementation-ready UI and interaction specification, zh-CN: 定义可直接实施的界面与交互规格 }
+outcomes: [Implementation-ready interface specification]
+artifacts: [ui-specification]
+inputs: [{ name: brief, type: text, required: true }]
+outputs: [{ name: specification, type: design-document, required: true }]
+preconditions: [Target users and primary task are identifiable]
+failures: [Critical product constraints are unavailable]
+sideEffects: [filesystem-write]
+permissions: { network: read, filesystem: workspace-write, shell: none, secrets: none, externalCommunication: none }
+taxonomy:
+  primaryCategory: design-media
+  lifecycle: [plan, create]
+  modalities: [image, text]
+  dependencies: [tool-agnostic]
+  risk: reversible-write
+routing:
+  positiveTriggers: { en: [design a user interface, create a UX specification, specify the screen], zh-CN: [设计用户界面, 创建交互规格, 设计页面] }
+  negativeTriggers: { en: [implement the frontend, analyze a data chart], zh-CN: [实现前端, 分析数据图表] }
+  confusableWith: [atom-implement-code-change]
+provenance: { origin: local, license: Apache-2.0 }
+evaluations: [routing-design, interface-spec-task]

package/skills/atom-find-research-sources/SKILL.md

@@ -0,0 +1,8 @@
+---
+name: atom-find-research-sources
+description: Find and rank authoritative sources for one research question with traceable links and relevance notes; use for evidence discovery, not for writing the final synthesis or inventing unsupported conclusions.
+---
+
+# Find Research Sources
+
+Clarify the claim or question, search authoritative and primary sources, record provenance and dates, remove irrelevant or duplicate results, and return a ranked source set with short relevance notes. Keep source discovery separate from final synthesis.

package/skills/atom-find-research-sources/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: Atom Find Research Sources
+  short_description: Route and apply Atom Find Research Sources
+  default_prompt: Use $atom-find-research-sources to handle this request within
+    its declared capability boundary.
+policy:
+  allow_implicit_invocation: true

package/skills/atom-find-research-sources/skill.contract.yaml

@@ -0,0 +1,26 @@
+schemaVersion: 1
+id: atom-find-research-sources
+kind: atom
+version: 0.1.0
+name: { en: Find research sources, zh-CN: 查找研究来源 }
+summary: { en: Return a ranked and traceable source set for one question, zh-CN: 为一个问题返回有排序且可追溯的来源集 }
+outcomes: [Ranked traceable research source set]
+artifacts: [source-list]
+inputs: [{ name: question, type: text, required: true }]
+outputs: [{ name: sources, type: citation-list, required: true }]
+preconditions: [Network research is authorized]
+failures: [No source meets the declared evidence standard]
+sideEffects: [network-read]
+permissions: { network: read, filesystem: none, shell: none, secrets: none, externalCommunication: none }
+taxonomy:
+  primaryCategory: research-knowledge
+  lifecycle: [discover, verify]
+  modalities: [text, structured-data]
+  dependencies: [api-backed]
+  risk: read-only
+routing:
+  positiveTriggers: { en: [find authoritative sources, research references, gather evidence], zh-CN: [查找权威来源, 搜集研究资料, 收集证据] }
+  negativeTriggers: { en: [write the final report, fabricate citations], zh-CN: [撰写最终报告, 编造引用] }
+  confusableWith: []
+provenance: { origin: local, license: Apache-2.0 }
+evaluations: [routing-research, source-quality-task]

package/skills/atom-implement-code-change/SKILL.md

@@ -0,0 +1,8 @@
+---
+name: atom-implement-code-change
+description: Implement one bounded, authorized software change and verify it in the current repository; use when source files should be edited, while keeping unrelated user changes intact and reporting evidence from tests.
+---
+
+# Implement a Code Change
+
+Confirm the requested behavior and repository constraints, inspect the smallest relevant surface, edit only authorized files, and verify in proportion to risk. Preserve unrelated changes. Report the changed behavior, files, and test evidence. Do not use for plan-only or security-audit-only requests.

package/skills/atom-implement-code-change/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: Atom Implement Code Change
+  short_description: Route and apply Atom Implement Code Change
+  default_prompt: Use $atom-implement-code-change to handle this request within
+    its declared capability boundary.
+policy:
+  allow_implicit_invocation: true

package/skills/atom-implement-code-change/skill.contract.yaml

@@ -0,0 +1,26 @@
+schemaVersion: 1
+id: atom-implement-code-change
+kind: atom
+version: 0.1.0
+name: { en: Implement a code change, zh-CN: 实现代码修改 }
+summary: { en: Edit and verify one bounded software behavior, zh-CN: 编辑并验证一个边界清晰的软件行为 }
+outcomes: [Verified bounded software change]
+artifacts: [source-change]
+inputs: [{ name: request, type: text, required: true }, { name: repository, type: workspace, required: true }]
+outputs: [{ name: change, type: diff, required: true }, { name: verification, type: test-evidence, required: true }]
+preconditions: [The requested edit is authorized, The repository is readable and writable]
+failures: [Requirements conflict with repository constraints, Verification cannot run]
+sideEffects: [filesystem-read, filesystem-write, shell-execution]
+permissions: { network: read, filesystem: workspace-write, shell: allowlisted, secrets: none, externalCommunication: none }
+taxonomy:
+  primaryCategory: software-engineering
+  lifecycle: [create, verify]
+  modalities: [code, system-state]
+  dependencies: [harness-backed]
+  risk: reversible-write
+routing:
+  positiveTriggers: { en: [implement the code change, implement a patch, fix this bug, edit the source], zh-CN: [实现代码修改, 实现补丁, 修复这个问题, 编辑源码] }
+  negativeTriggers: { en: [plan only, security audit only], zh-CN: [只做计划, 只做安全审计] }
+  confusableWith: [atom-plan-code-change, atom-audit-source-security]
+provenance: { origin: local, license: Apache-2.0 }
+evaluations: [routing-bootstrap, implement-code-task]

package/skills/atom-organize-personal-tasks/SKILL.md

@@ -0,0 +1,8 @@
+---
+name: atom-organize-personal-tasks
+description: Turn one person's unstructured commitments into a prioritized, dated, and reviewable task list without sending notifications; use for personal organization rather than team workflow automation.
+---
+
+# Organize Personal Tasks
+
+Extract commitments, clarify outcomes and deadlines, separate projects from next actions, identify dependencies, prioritize transparently, and return a reviewable task list. Do not create external tasks, calendar events, or reminders unless a separate authorized write atom is selected.

package/skills/atom-organize-personal-tasks/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: Atom Organize Personal Tasks
+  short_description: Route and apply Atom Organize Personal Tasks
+  default_prompt: Use $atom-organize-personal-tasks to handle this request within
+    its declared capability boundary.
+policy:
+  allow_implicit_invocation: true

package/skills/atom-organize-personal-tasks/skill.contract.yaml

@@ -0,0 +1,26 @@
+schemaVersion: 1
+id: atom-organize-personal-tasks
+kind: atom
+version: 0.1.0
+name: { en: Organize personal tasks, zh-CN: 整理个人任务 }
+summary: { en: Convert unstructured commitments into a prioritized personal task list, zh-CN: 把零散承诺整理成有优先级的个人任务清单 }
+outcomes: [Prioritized personal task list]
+artifacts: [task-list]
+inputs: [{ name: commitments, type: text, required: true }]
+outputs: [{ name: tasks, type: structured-task-list, required: true }]
+preconditions: [The commitments belong to one person's scope]
+failures: [Essential deadlines or ownership remain contradictory]
+sideEffects: [none]
+permissions: { network: none, filesystem: none, shell: none, secrets: none, externalCommunication: none }
+taxonomy:
+  primaryCategory: personal-productivity
+  lifecycle: [plan]
+  modalities: [text, structured-data]
+  dependencies: [tool-agnostic]
+  risk: read-only
+routing:
+  positiveTriggers: { en: [organize my tasks, prioritize my todo list, turn notes into next actions], zh-CN: [整理我的任务, 安排待办优先级, 把笔记变成行动项] }
+  negativeTriggers: { en: [automate the team workflow, send reminders now], zh-CN: [自动化团队工作流, 立即发送提醒] }
+  confusableWith: [atom-design-automation-workflow]
+provenance: { origin: local, license: Apache-2.0 }
+evaluations: [routing-productivity, task-organization-task]

package/skills/atom-plan-code-change/SKILL.md

@@ -0,0 +1,8 @@
+---
+name: atom-plan-code-change
+description: Produce an implementation-ready plan for one bounded software change without editing files; use when the user wants scope, affected components, risks, tests, and ordered steps before implementation.
+---
+
+# Plan a Code Change
+
+Inspect the relevant repository evidence, define the behavioral change and non-goals, identify exact files or components, and produce ordered implementation and verification steps. Call out unknowns and risks. Do not edit source files; requests that include implementation belong to `atom-implement-code-change` or a software-delivery pack.

package/skills/atom-plan-code-change/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: Atom Plan Code Change
+  short_description: Route and apply Atom Plan Code Change
+  default_prompt: Use $atom-plan-code-change to handle this request within its
+    declared capability boundary.
+policy:
+  allow_implicit_invocation: true

package/skills/atom-plan-code-change/skill.contract.yaml

@@ -0,0 +1,26 @@
+schemaVersion: 1
+id: atom-plan-code-change
+kind: atom
+version: 0.1.0
+name: { en: Plan a code change, zh-CN: 规划代码修改 }
+summary: { en: Produce an implementation-ready plan without changing source files, zh-CN: 在不修改源码的前提下产出可实施计划 }
+outcomes: [Implementation-ready code change plan]
+artifacts: [plan]
+inputs: [{ name: request, type: text, required: true }, { name: repository, type: workspace, required: true }]
+outputs: [{ name: plan, type: markdown, required: true }]
+preconditions: [Relevant repository content is readable]
+failures: [The requested behavior cannot be located, Critical requirements remain contradictory]
+sideEffects: [filesystem-read]
+permissions: { network: none, filesystem: read, shell: allowlisted, secrets: none, externalCommunication: none }
+taxonomy:
+  primaryCategory: software-engineering
+  lifecycle: [plan]
+  modalities: [code, text]
+  dependencies: [harness-backed]
+  risk: read-only
+routing:
+  positiveTriggers: { en: [plan a code change, implementation plan, inspect and plan], zh-CN: [规划代码修改, 实施计划, 先分析再规划] }
+  negativeTriggers: { en: [implement the change, edit the files], zh-CN: [实现修改, 编辑文件] }
+  confusableWith: [atom-implement-code-change]
+provenance: { origin: local, license: Apache-2.0 }
+evaluations: [routing-bootstrap, plan-code-task]

package/skills/atom-translate-document/SKILL.md

@@ -0,0 +1,8 @@
+---
+name: atom-translate-document
+description: Translate one human-readable document into a requested language while preserving meaning, structure, terminology, and unresolved ambiguities; use for translation, not source-code localization or unrelated rewriting.
+---
+
+# Translate a Document
+
+Identify audience, locale, terminology, and formatting constraints; translate the complete content; preserve headings, relationships, citations, and placeholders; flag genuine ambiguities; and verify that no section was omitted. Do not silently summarize or alter the source claims.

package/skills/atom-translate-document/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: Atom Translate Document
+  short_description: Route and apply Atom Translate Document
+  default_prompt: Use $atom-translate-document to handle this request within its
+    declared capability boundary.
+policy:
+  allow_implicit_invocation: true

package/skills/atom-translate-document/skill.contract.yaml

@@ -0,0 +1,26 @@
+schemaVersion: 1
+id: atom-translate-document
+kind: atom
+version: 0.1.0
+name: { en: Translate a document, zh-CN: 翻译文档 }
+summary: { en: Translate one document while preserving meaning and structure, zh-CN: 在保持含义与结构的前提下翻译一份文档 }
+outcomes: [Meaning-preserving translated document]
+artifacts: [translated-document]
+inputs: [{ name: document, type: document, required: true }, { name: target-locale, type: locale, required: true }]
+outputs: [{ name: translation, type: document, required: true }]
+preconditions: [The source document is readable]
+failures: [Critical terminology cannot be resolved, The source is incomplete]
+sideEffects: [filesystem-read, filesystem-write]
+permissions: { network: none, filesystem: workspace-write, shell: none, secrets: none, externalCommunication: none }
+taxonomy:
+  primaryCategory: documents-communication
+  lifecycle: [transform, verify]
+  modalities: [document, text]
+  dependencies: [tool-agnostic]
+  risk: reversible-write
+routing:
+  positiveTriggers: { en: [translate this document, localize this report], zh-CN: [翻译这份文档, 本地化这份报告] }
+  negativeTriggers: { en: [translate source code, summarize only], zh-CN: [翻译源代码, 只做摘要] }
+  confusableWith: []
+provenance: { origin: local, license: Apache-2.0 }
+evaluations: [routing-bootstrap, translation-task]

package/skills/category-automation-operations/SKILL.md

@@ -0,0 +1,13 @@
+---
+name: category-automation-operations
+description: Route workflow integration, scheduling, deployment, monitoring, infrastructure, and repeatable operational requests to focused Atomic Skills; use when systems must be connected or operated rather than merely designed.
+---
+
+# Automation and Operations Router
+
+1. Identify triggers, systems, credentials, side effects, recovery, and observability needs.
+2. Read the localized Atomic Skill index in `references/`.
+3. Separate workflow design from deployment and live operation when their permissions differ.
+4. Prefer reversible, idempotent atoms and declare all MCP/API dependencies.
+
+Do not hide one-off product implementation or personal reminders inside this category.

package/skills/category-automation-operations/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: Category Automation Operations
+  short_description: Route and apply Category Automation Operations
+  default_prompt: Use $category-automation-operations to handle this request
+    within its declared capability boundary.
+policy:
+  allow_implicit_invocation: true

package/skills/category-automation-operations/references/index.en.md

@@ -0,0 +1,11 @@
+# Automation and operations: Atomic Skills
+
+Connect systems, schedule work, deploy services, and operate repeatable workflows.
+
+## Atomic Skills
+
+- `atom-design-automation-workflow` — **Design an automation workflow**: Specify one safe and observable repeatable workflow
+  - Risk: `reversible-write`
+  - Lifecycle: plan
+
+Read only the selected Atomic Skill. Use a capability pack when the request has independently useful outcomes.

package/skills/category-automation-operations/references/index.md

@@ -0,0 +1,11 @@
+# Automation and operations: Atomic Skills
+
+Connect systems, schedule work, deploy services, and operate repeatable workflows.
+
+## Atomic Skills
+
+- `atom-design-automation-workflow` — **Design an automation workflow**: Specify one safe and observable repeatable workflow
+  - Risk: `reversible-write`
+  - Lifecycle: plan
+
+Read only the selected Atomic Skill. Use a capability pack when the request has independently useful outcomes.

package/skills/category-automation-operations/references/index.zh-CN.md

@@ -0,0 +1,11 @@
+# 自动化与运营: 原子 Skill
+
+连接系统、调度工作、部署服务并运行可重复流程。
+
+## 原子 Skill
+
+- `atom-design-automation-workflow` — **设计自动化工作流**: 规定一个安全、可观测的可重复工作流
+  - 风险: `reversible-write`
+  - 生命周期: plan
+
+只读取最终选中的原子 Skill；若请求包含多个独立结果，请改用能力包。

package/skills/category-automation-operations/references/index.zh.md

@@ -0,0 +1,11 @@
+# 自动化与运营: 原子 Skill
+
+连接系统、调度工作、部署服务并运行可重复流程。
+
+## 原子 Skill
+
+- `atom-design-automation-workflow` — **设计自动化工作流**: 规定一个安全、可观测的可重复工作流
+  - 风险: `reversible-write`
+  - 生命周期: plan
+
+只读取最终选中的原子 Skill；若请求包含多个独立结果，请改用能力包。

package/skills/category-automation-operations/skill.contract.yaml

@@ -0,0 +1,26 @@
+schemaVersion: 1
+id: category-automation-operations
+kind: category
+version: 0.1.0
+name: { en: Automation and Operations Router, zh-CN: 自动化与运营路由器 }
+summary: { en: Route repeatable system operation and integration work to focused atoms., zh-CN: 将可重复的系统运行与集成工作路由到聚焦原子能力。 }
+outcomes: [Bounded automation-operations category decision]
+artifacts: [category-route, localized-index]
+inputs: [{ name: request, type: user-prompt, required: true }]
+outputs: [{ name: candidates, type: category-and-atom-candidates, required: true }]
+preconditions: [The versioned taxonomy and localized index are available]
+failures: [The operational trigger or system boundary is unclear]
+sideEffects: [filesystem-read]
+permissions: { network: none, filesystem: read, shell: none, secrets: none, externalCommunication: none }
+taxonomy:
+  primaryCategory: automation-operations
+  lifecycle: [analyze]
+  modalities: [text, structured-data]
+  dependencies: [tool-agnostic]
+  risk: read-only
+routing:
+  positiveTriggers: { en: [design automation, schedule a workflow, deploy a service, monitor operations], zh-CN: [设计自动化, 调度工作流, 部署服务, 监控运维] }
+  negativeTriggers: { en: [one-off code change, personal reminder], zh-CN: [一次性代码修改, 个人提醒] }
+  confusableWith: [category-software-engineering, category-personal-productivity]
+provenance: { origin: local, license: Apache-2.0 }
+evaluations: [routing-bootstrap, routing-en-test, routing-zh-cn-test, routing-adversarial]

package/skills/category-business-growth/SKILL.md

@@ -0,0 +1,13 @@
+---
+name: category-business-growth
+description: Route market, customer, sales, marketing, SEO, finance, investment, and growth decisions to bounded Atomic Skills; use when a business decision or commercial outcome is primary and domain-specific evidence is required.
+---
+
+# Business and Growth Router
+
+1. Identify the decision, market, time horizon, evidence standard, and risk level.
+2. Read the locale-appropriate Atomic Skill index in `references/`.
+3. Compose research or data atoms explicitly instead of duplicating them inside broad marketing Skills.
+4. Escalate financial, legal, or external-communication authority through the declared risk envelope.
+
+Do not use this category for generic research without a business decision.

package/skills/category-business-growth/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: Category Business Growth
+  short_description: Route and apply Category Business Growth
+  default_prompt: Use $category-business-growth to handle this request within its
+    declared capability boundary.
+policy:
+  allow_implicit_invocation: true

package/skills/category-business-growth/references/index.en.md

@@ -0,0 +1,11 @@
+# Business and growth: Atomic Skills
+
+Improve markets, sales, marketing, finance, and business decisions.
+
+## Atomic Skills
+
+- `atom-analyze-market` — **Analyze a market**: Produce evidence-backed findings for one commercial market question
+  - Risk: `read-only`
+  - Lifecycle: discover, analyze
+
+Read only the selected Atomic Skill. Use a capability pack when the request has independently useful outcomes.

package/skills/category-business-growth/references/index.md

@@ -0,0 +1,11 @@
+# Business and growth: Atomic Skills
+
+Improve markets, sales, marketing, finance, and business decisions.
+
+## Atomic Skills
+
+- `atom-analyze-market` — **Analyze a market**: Produce evidence-backed findings for one commercial market question
+  - Risk: `read-only`
+  - Lifecycle: discover, analyze
+
+Read only the selected Atomic Skill. Use a capability pack when the request has independently useful outcomes.

package/skills/category-business-growth/references/index.zh-CN.md

@@ -0,0 +1,11 @@
+# 商业与增长: 原子 Skill
+
+改善市场、销售、营销、财务和商业决策。
+
+## 原子 Skill
+
+- `atom-analyze-market` — **分析市场**: 为一个商业市场问题产出有证据的分析
+  - 风险: `read-only`
+  - 生命周期: discover, analyze
+
+只读取最终选中的原子 Skill；若请求包含多个独立结果，请改用能力包。

package/skills/category-business-growth/references/index.zh.md

@@ -0,0 +1,11 @@
+# 商业与增长: 原子 Skill
+
+改善市场、销售、营销、财务和商业决策。
+
+## 原子 Skill
+
+- `atom-analyze-market` — **分析市场**: 为一个商业市场问题产出有证据的分析
+  - 风险: `read-only`
+  - 生命周期: discover, analyze
+
+只读取最终选中的原子 Skill；若请求包含多个独立结果，请改用能力包。

package/skills/category-business-growth/skill.contract.yaml

@@ -0,0 +1,26 @@
+schemaVersion: 1
+id: category-business-growth
+kind: category
+version: 0.1.0
+name: { en: Business and Growth Router, zh-CN: 商业与增长路由器 }
+summary: { en: Route commercial decisions to bounded market and growth atoms., zh-CN: 将商业决策路由到有边界的市场与增长原子能力。 }
+outcomes: [Bounded business-growth category decision]
+artifacts: [category-route, localized-index]
+inputs: [{ name: request, type: user-prompt, required: true }]
+outputs: [{ name: candidates, type: category-and-atom-candidates, required: true }]
+preconditions: [The versioned taxonomy and localized index are available]
+failures: [No business decision or commercial outcome can be identified]
+sideEffects: [filesystem-read]
+permissions: { network: none, filesystem: read, shell: none, secrets: none, externalCommunication: none }
+taxonomy:
+  primaryCategory: business-growth
+  lifecycle: [analyze]
+  modalities: [text, structured-data]
+  dependencies: [tool-agnostic]
+  risk: read-only
+routing:
+  positiveTriggers: { en: [analyze a market, plan marketing, assess sales, evaluate growth], zh-CN: [分析市场, 规划营销, 评估销售, 评估增长] }
+  negativeTriggers: { en: [generic research, personal task tracking], zh-CN: [通用研究, 个人任务管理] }
+  confusableWith: [category-research-knowledge, category-data-analytics]
+provenance: { origin: local, license: Apache-2.0 }
+evaluations: [routing-bootstrap, routing-en-test, routing-zh-cn-test, routing-adversarial]

package/skills/category-data-analytics/SKILL.md

@@ -0,0 +1,13 @@
+---
+name: category-data-analytics
+description: Route structured-data transformation, querying, calculation, statistical analysis, modeling, and charting to focused Atomic Skills; use when the correctness of a dataset, metric, model, or data-derived result is primary.
+---
+
+# Data and Analytics Router
+
+1. Identify the data shape, analytical question, and required reproducible output.
+2. Read the localized Atomic Skill index in `references/`.
+3. Distinguish analysis from the later narrative report or visual brand treatment.
+4. Select atoms with compatible input/output contracts and risk boundaries.
+
+Do not use this category for prose-only research or production service operations.

package/skills/category-data-analytics/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: Category Data Analytics
+  short_description: Route and apply Category Data Analytics
+  default_prompt: Use $category-data-analytics to handle this request within its
+    declared capability boundary.
+policy:
+  allow_implicit_invocation: true

package/skills/category-data-analytics/references/index.en.md

@@ -0,0 +1,11 @@
+# Data and analytics: Atomic Skills
+
+Transform, query, model, calculate, or visualize structured data.
+
+## Atomic Skills
+
+- `atom-analyze-tabular-data` — **Analyze tabular data**: Answer one structured-data question with reproducible calculations
+  - Risk: `read-only`
+  - Lifecycle: analyze, verify
+
+Read only the selected Atomic Skill. Use a capability pack when the request has independently useful outcomes.

package/skills/category-data-analytics/references/index.md

@@ -0,0 +1,11 @@
+# Data and analytics: Atomic Skills
+
+Transform, query, model, calculate, or visualize structured data.
+
+## Atomic Skills
+
+- `atom-analyze-tabular-data` — **Analyze tabular data**: Answer one structured-data question with reproducible calculations
+  - Risk: `read-only`
+  - Lifecycle: analyze, verify
+
+Read only the selected Atomic Skill. Use a capability pack when the request has independently useful outcomes.

package/skills/category-data-analytics/references/index.zh-CN.md

@@ -0,0 +1,11 @@
+# 数据与分析: 原子 Skill
+
+转换、查询、建模、计算或可视化结构化数据。
+
+## 原子 Skill
+
+- `atom-analyze-tabular-data` — **分析表格数据**: 用可复现计算回答一个结构化数据问题
+  - 风险: `read-only`
+  - 生命周期: analyze, verify
+
+只读取最终选中的原子 Skill；若请求包含多个独立结果，请改用能力包。

package/skills/category-data-analytics/references/index.zh.md

@@ -0,0 +1,11 @@
+# 数据与分析: 原子 Skill
+
+转换、查询、建模、计算或可视化结构化数据。
+
+## 原子 Skill
+
+- `atom-analyze-tabular-data` — **分析表格数据**: 用可复现计算回答一个结构化数据问题
+  - 风险: `read-only`
+  - 生命周期: analyze, verify
+
+只读取最终选中的原子 Skill；若请求包含多个独立结果，请改用能力包。

package/skills/category-data-analytics/skill.contract.yaml

@@ -0,0 +1,26 @@
+schemaVersion: 1
+id: category-data-analytics
+kind: category
+version: 0.1.0
+name: { en: Data and Analytics Router, zh-CN: 数据与分析路由器 }
+summary: { en: Route structured-data questions to reproducible analysis atoms., zh-CN: 将结构化数据问题路由到可复现的分析原子能力。 }
+outcomes: [Bounded data-analytics category decision]
+artifacts: [category-route, localized-index]
+inputs: [{ name: request, type: user-prompt, required: true }]
+outputs: [{ name: candidates, type: category-and-atom-candidates, required: true }]
+preconditions: [The versioned taxonomy and localized index are available]
+failures: [The data shape or analytical outcome cannot be inferred]
+sideEffects: [filesystem-read]
+permissions: { network: none, filesystem: read, shell: none, secrets: none, externalCommunication: none }
+taxonomy:
+  primaryCategory: data-analytics
+  lifecycle: [analyze]
+  modalities: [text, structured-data]
+  dependencies: [tool-agnostic]
+  risk: read-only
+routing:
+  positiveTriggers: { en: [analyze a table, query data, calculate metrics, build a statistical model], zh-CN: [分析表格, 查询数据, 计算指标, 建立统计模型] }
+  negativeTriggers: { en: [prose-only research, operate a production service], zh-CN: [纯文字研究, 运行生产服务] }
+  confusableWith: [category-research-knowledge, category-design-media]
+provenance: { origin: local, license: Apache-2.0 }
+evaluations: [routing-bootstrap, routing-en-test, routing-zh-cn-test, routing-adversarial]

package/skills/category-design-media/SKILL.md

@@ -0,0 +1,13 @@
+---
+name: category-design-media
+description: Route user-interface specification, visual design, diagramming, image, audio, and video creation or editing to modality-specific Atomic Skills; use when perceptual quality or media transformation is the primary outcome.
+---
+
+# Design and Media Router
+
+1. Identify the medium, audience, design constraints, and required editable artifact.
+2. Read the localized Atomic Skill index in `references/`.
+3. Distinguish interface design from frontend code implementation and data charts from illustration.
+4. Select the least-privileged atom set and retain source asset provenance.
+
+Do not use this category when executable code or quantitative correctness is the primary result.

package/skills/category-design-media/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: Category Design Media
+  short_description: Route and apply Category Design Media
+  default_prompt: Use $category-design-media to handle this request within its
+    declared capability boundary.
+policy:
+  allow_implicit_invocation: true

package/skills/category-design-media/references/index.en.md

@@ -0,0 +1,11 @@
+# Design and media: Atomic Skills
+
+Design interfaces or create and transform visual, audio, and video media.
+
+## Atomic Skills
+
+- `atom-design-interface-spec` — **Design an interface specification**: Define an implementation-ready UI and interaction specification
+  - Risk: `reversible-write`
+  - Lifecycle: plan, create
+
+Read only the selected Atomic Skill. Use a capability pack when the request has independently useful outcomes.

package/skills/category-design-media/references/index.md

@@ -0,0 +1,11 @@
+# Design and media: Atomic Skills
+
+Design interfaces or create and transform visual, audio, and video media.
+
+## Atomic Skills
+
+- `atom-design-interface-spec` — **Design an interface specification**: Define an implementation-ready UI and interaction specification
+  - Risk: `reversible-write`
+  - Lifecycle: plan, create
+
+Read only the selected Atomic Skill. Use a capability pack when the request has independently useful outcomes.

package/skills/category-design-media/references/index.zh-CN.md

@@ -0,0 +1,11 @@
+# 设计与媒体: 原子 Skill
+
+设计界面或创建、转换图像、音频和视频媒体。
+
+## 原子 Skill
+
+- `atom-design-interface-spec` — **设计界面规格**: 定义可直接实施的界面与交互规格
+  - 风险: `reversible-write`
+  - 生命周期: plan, create
+
+只读取最终选中的原子 Skill；若请求包含多个独立结果，请改用能力包。