concept-skills
Health Gecti
- License — License: Apache-2.0
- Description — Repository has a description
- Active repo — Last push 0 days ago
- Community trust — 13 GitHub stars
Code Gecti
- Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Gecti
- Permissions — No dangerous permissions requested
Bu listing icin henuz AI raporu yok.
Agent Skills for concept-driven software design, semantic modeling, ontology, knowledge management, and visual learning. English and Simplified Chinese editions for Claude Code, Codex, and Cursor.
concept-skills
Name the meaning first — then write code, run a CLI, or draft a spec.
concept-skills provides eleven
capabilities as 22 localized Agent Skills for ontology, semantic layers,
concept design, and visual learning. English uses the base name;
Simplified Chinese adds -cn.
Cloud operational skills now live inconcept-git/cloud-concept-skills and are no
longer distributed from this repository.
Office-focused skills now live in the sibling myoffice-skills repository and are no longer
distributed from this repository.
The software-concept-architect-design skills adapt Daniel Jackson's concepts-and-synchronizations model —
The Essence of Software (2021), with the current when/where/then
sync notation from Beyond Objects (arXiv:2606.27258) — for
agent use; an adaptation, not endorsed by the author.
中文说明见 README-CN.md.
Current release is 2.4.0. The concept design workflow, see-the-concept, and settle-the-concept-design are explicit-only;software-concept-architect-refine and read-it-reframe-it-own-it support automatic discovery;software-concept-architect-guardrails modes
consume Jackson notation only.
Table of Contents
Skills
Classified by the primary problem each skill solves; family names identify the series.
Concept Design
Design software concepts, responsibilities, specifications, and implementations.
| Skill | Version | ClawHub category |
|---|---|---|
| Settle the Concept Design | 0.1.1 | development |
| Software Concept Architect · Design | 1.0.0 | development |
| Software Concept Architect · PRD | 1.1.0 | development |
| Software Concept Architect · Build | 1.1.0 | development |
| Software Concept Architect · Review | 1.1.0 | development |
| Software Concept Architect · Guardrails | 1.1.0 | development |
| Software Concept Architect · Refine | 1.0.0 | development |
| See the Concept | 0.1.4 | knowledge |
Semantic Modeling
Model the business meaning of data, facts, dimensions, and relations.
| Skill | Version | ClawHub category |
|---|---|---|
| Data Knowledge Architect | 1.1.0 | knowledge |
Knowledge Management
Extract, connect, organize, and reuse knowledge from source material.
| Skill | Version | ClawHub category |
|---|---|---|
| Personal Knowledge Architect | 0.4.0 | knowledge |
Learning Methods
Build and revise understanding around a learning goal and a key framework.
| Skill | Version | ClawHub category |
|---|---|---|
| Read It. Reframe It. Own It. | 0.2.0 | knowledge |
Concept design workflow
Purpose-Driven Software Design. Based on Daniel Jackson’s concept design approach in The Essence of Software.
Requirements to modules, with the concept model as the contract. Each skill stops at its own
boundary and hands off: design → prd / build → review.
Shared principle: synchronization may restrict behavior, never extend a concept contract. Check
contract conformance and purpose fulfillment separately.
Per-skill details and taxonomy: catalog.yml.
Localization and directory rules: CLAUDE.md.
Install
Requires Node.js 22.20+ and Git. Copy a command below and run it from your project directory. These commands install to Codex in the current project.
Use --agent claude-code or --agent cursor for another agent; add --global to install across projects. See the installer options.
English editions
npx --yes skills add ontology-of-everything/concept-skills --skill data-knowledge-architect personal-knowledge-architect software-concept-architect-design software-concept-architect-prd software-concept-architect-build software-concept-architect-review software-concept-architect-guardrails software-concept-architect-refine read-it-reframe-it-own-it see-the-concept settle-the-concept-design --agent codex --copy --yes
Simplified Chinese editions
npx --yes skills add ontology-of-everything/concept-skills --skill data-knowledge-architect-cn personal-knowledge-architect-cn software-concept-architect-design-cn software-concept-architect-prd-cn software-concept-architect-build-cn software-concept-architect-review-cn software-concept-architect-guardrails-cn software-concept-architect-refine-cn read-it-reframe-it-own-it-cn see-the-concept-cn settle-the-concept-design-cn --agent codex --copy --yes
Install one skill
data-knowledge-architect
npx --yes skills add ontology-of-everything/concept-skills --skill data-knowledge-architect --agent codex --copy --yes
personal-knowledge-architect
npx --yes skills add ontology-of-everything/concept-skills --skill personal-knowledge-architect --agent codex --copy --yes
software-concept-architect-design
npx --yes skills add ontology-of-everything/concept-skills --skill software-concept-architect-design --agent codex --copy --yes
software-concept-architect-prd
npx --yes skills add ontology-of-everything/concept-skills --skill software-concept-architect-prd --agent codex --copy --yes
software-concept-architect-build
npx --yes skills add ontology-of-everything/concept-skills --skill software-concept-architect-build --agent codex --copy --yes
software-concept-architect-review
npx --yes skills add ontology-of-everything/concept-skills --skill software-concept-architect-review --agent codex --copy --yes
software-concept-architect-guardrails
npx --yes skills add ontology-of-everything/concept-skills --skill software-concept-architect-guardrails --agent codex --copy --yes
software-concept-architect-refine
npx --yes skills add ontology-of-everything/concept-skills --skill software-concept-architect-refine --agent codex --copy --yes
read-it-reframe-it-own-it
npx --yes skills add ontology-of-everything/concept-skills --skill read-it-reframe-it-own-it --agent codex --copy --yes
see-the-concept
npx --yes skills add ontology-of-everything/concept-skills --skill see-the-concept --agent codex --copy --yes
List available skills
npx --yes skills add ontology-of-everything/concept-skills --list
English payloads live in skills/en/; Simplified Chinese payloads live in skills/zh-CN/ and keep the -cn skill names. ClawHub publishes English editions only.
Usage
Skills normally activate from their description, so plain requests are enough. The
concept design workflow is explicit-only and must be named (/software-concept-architect-design in Cursor,$software-concept-architect-design in Codex):
$software-concept-architect-design model this requirement as independent concepts
$software-concept-architect-prd transcribe the confirmed model into a PRD
$software-concept-architect-guardrails drift src/orders
$see-the-concept explain this concept model with a small diagram
Turn this API into a semantic layer → data-knowledge-architect
To pin a skill explicitly: /skill-name in Cursor, $skill-name in Codex.
Classic usage examples
Each skill has two examples. Install it, share the relevant material, then copy a request.
Replace example paths with your project paths.
Settle the Concept Design
settle-the-concept-design · Skill
Based on Daniel Jackson's concept design, consolidate discussion into decisions for confirmation, from goals and concept purposes to scenarios and the specifications needed for those decisions.
| Scenario | Try saying |
|---|---|
| Settle a design | $settle-the-concept-design Consolidate our discussion into a concept design; show the choices, rationale, and unresolved issues for confirmation. |
| Revisit decisions | $settle-the-concept-design Compare this round with our confirmed design; show what changed, why, and how it affects business goals and acceptance criteria. |
Read It. Reframe It. Own It
read-it-reframe-it-own-it
| Scenario | Try saying |
|---|---|
| Organize notes | $read-it-reframe-it-own-it Turn these concept-design notes into a knowledge tree I can use for revision. |
| Make a study sketchnote | $read-it-reframe-it-own-it Turn this knowledge tree into a Chinese study sketchnote. |
The second example also needs an image-generation tool. The English skill is on ClawHub;
its marketplace URL uses a different slug from the repository installation name.
See the Concept
see-the-concept · Usage and visual toolbox
Help beginners understand and examine software concepts and how they work together, from overview to detail, using source evidence and visuals. The bundled design reference provides reusable HTML, CSS, and JavaScript.
| Scenario | Try saying |
|---|---|
| Understand a design | $see-the-concept Explore this project and visualize concept boundaries and coordinated actions. |
| Check implementation | $see-the-concept Compare this design with source and explain differences and their effects. |
Personal Knowledge Architect
personal-knowledge-architect
| Scenario | Try saying |
|---|---|
| Find reusable methods | $personal-knowledge-architect Pull out reusable methods from these interview notes. Show me the outline first. |
| Build a method library | $personal-knowledge-architect Turn these product-research notes into a method library and merge duplicates. |
Data Knowledge Architect
data-knowledge-architect
| Scenario | Try saying |
|---|---|
| Understand a table | $data-knowledge-architect Which parts of this order table are facts, dimensions, and measures? |
| Model APIs | $data-knowledge-architect Build a semantic model from these order APIs. Export OKF after we confirm it. |
Software Concept Architect · Design
software-concept-architect-design
| Scenario | Try saying |
|---|---|
| Design a task list | $software-concept-architect-design Help me design a task list that supports adding, completing, and reopening tasks. |
| Design reservations | $software-concept-architect-design When a restaurant is full, customers can join a waiting list. What concepts do we need, and how do they work together? |
Software Concept Architect · PRD
software-concept-architect-prd
| Scenario | Try saying |
|---|---|
| Write a PRD | $software-concept-architect-prd Turn our confirmed task-list model into a short PRD. |
| Write specifications | $software-concept-architect-prd Write CONCEPT.md and SYNCS.md for this reservation system using our confirmed design. |
Software Concept Architect · Build
software-concept-architect-build
| Scenario | Try saying |
|---|---|
| Build a task list | $software-concept-architect-build Build our confirmed task-list model in TypeScript and test it. |
| Build reservations | $software-concept-architect-build Implement reservations and cancellations from these specs, then check the flow. |
Software Concept Architect · Review
software-concept-architect-review
| Scenario | Try saying |
|---|---|
| Review a model | $software-concept-architect-review Check this task-list model for problems. Leave it unchanged for now. |
| Compare code with specs | $software-concept-architect-review Check the reservation code against its specs and show me where they differ. |
Software Concept Architect · Guardrails
software-concept-architect-guardrails
| Scenario | Try saying |
|---|---|
| Find missing specs | $software-concept-architect-guardrails audit src/reservations Which specs are missing here? List them for me. |
| Check drift | $software-concept-architect-guardrails drift src Check whether the code matches its specs. Report only. |
Software Concept Architect · Refine
software-concept-architect-refine
| Scenario | Try saying |
|---|---|
| Check responsibilities | $software-concept-architect-refine This Todo handles tasks and sends reminders. Should we split it? |
| Compare similar concepts | $software-concept-architect-refine Should Favorites and Read Later stay separate? Walk me through the use cases. |
A connected workflow: design and confirm the model → PRD → implementation → review. For an existing project, start with Guardrails audit, then backfill specifications or refine concepts as needed.
Use see-the-concept at any step to explain the current model or implementation visually.
Contributing
qa/ is private and untracked; public checks do not depend on it. Run ./tools/validate-all.sh --local-qa only when local QA is available.
Project rules and repository ownership: CLAUDE.md.
git config core.hooksPath .githooks
./tools/validate-all.sh
skills/en/<name>/ and skills/zh-CN/<name>-cn/ are self-contained install payloads. QA stays private and untracked. Update both locales, the public catalog, README views, and parity checks together.
License
Apache-2.0 © concept-skills contributors. Bundles published to
ClawHub are MIT-0 there; the repository source stays Apache-2.0.
Yorumlar (0)
Yorum birakmak icin giris yap.
Yorum birakSonuc bulunamadi