Domain glossary
My personal Claude Code skills. Engineering process, context hygiene, and AI-agent maintenance — straight from my .claude directory.
npx -y skills add silvio-l/skills --skill domain-glossaryAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Build and maintain the project's domain vocabulary, glossary, and key conceptual decisions in CONTEXT.md. Use when the user wants to create, review, or refine project terminology, glossary entries, or conceptual decisions in CONTEXT.md.
SKILL.md
3.6 KB, as published. Nobody here has run it
Domain Glossary
Work with the user to create or improve the project's domain vocabulary and key conceptual decisions. CONTEXT.md is the storage vehicle — the real work is getting the language right.
The target CONTEXT.md is the one in the current project's root — not this skills-source repo's CONTEXT.md. If invoked inside the silvio-l/skills repo itself, treat the existing CONTEXT.md as a skill-authoring meta-glossary (not a business domain) and confirm with the user before extending it.
Start
- Check whether CONTEXT.md exists in the project root.
- If it exists: read it fully. Identify existing terms, definitions, conventions, and conceptual decisions. Flag duplicates, contradictions, vague formulations, and potentially stale content — but propose changes only, change nothing yet.
- If it does not exist: propose a structure and section outline. Ask for approval before writing anything.
Mandatory: use the grill-me skill throughout
Invoke the grill-me skill for every relevant term or conceptual decision:
- Challenge each term: is it precise, unambiguous, future-proof, AI-readable?
- Ask about scope: what does this term include? What does it explicitly exclude?
- Surface alternatives and explain the trade-offs.
- Point out consequences of a choice (e.g. naming collision with a library, ambiguity with another domain term).
- Never silently accept vague or overloaded language.
Ask questions one at a time.
Step-by-step workflow
1. Analyze existing content (read-only)
2. Collect terms and open questions
3. Mark unclear, duplicate, or contradictory items
4. Ask targeted questions (one at a time)
5. Present options with recommendation
6. Wait for explicit approval
7. Only then: write to CONTEXT.md
8. After writing: re-read the modified section and show it to the user to confirm it landed correctly
Decision format
For every term or formulation in question, present:
| Field | Content |
|---|---|
| Current | existing term or phrasing (or "none") |
| Problem | what is unclear, duplicate, contradictory, or improvable |
| Options | 2–3 concrete alternatives |
| Recommendation | your preferred option and why |
| Decision needed | explicit yes/no question for the user |
Language
The CONTEXT.md artifact is authored in German — terms, definitions, decisions, explanations. This is a deliberate domain choice, not a default: the project's domain language is German, so the glossary must be too. Technical identifiers (class, method, field names) stay in their original form (usually English) but are explained in German. Never switch to English, even when an English phrasing feels more precise — find the German equivalent instead and carry the precision through context.
Hard rules — never bypass
The following actions require explicit user approval before execution:
- Introduce a new term
- Rename an existing term
- Change a definition
- Remove content from CONTEXT.md
- Merge sections
- Mark something as a binding project convention
- Mark existing content as outdated
- Save any file changes
Make no silent assumptions. Only document what has been explicitly confirmed.
Goal
A CONTEXT.md that is clear, consistent, and AI-readable — where the user retains full conscious control over language, meaning, and conceptual decisions. The AI structures, questions, and challenges. The user decides.