OpenParable
Skip to content
← Back to skills

Glossary

Documents
glossary

Glossary is an Agent Skill for documents published in the majiayu000/claude-skill-registry repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Build or update a CONTEXT.md domain glossary for the current project. Establishes shared vocabulary with term definitions, avoid-aliases, and relationships so agents use consistent, concise language. This skill should be used when starting a new project, when domain terminology is inconsistent across sessions, when the user says "build a glossary", "define our terms", "shared vocabulary", or "domain language", or when an existing CONTEXT.md needs updating after new domain concepts emerge.

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/majiayu000/claude-skill-registry/blob/HEAD/skills/context-management/glossary/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/glossary, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Glossary — Domain Language Builder

Build or update a CONTEXT.md at the project root that captures the project's domain language. A shared vocabulary between user and agent reduces verbosity (agents stop using 20 words where 1 domain term suffices), enforces consistent naming in code, and prevents terminology drift across sessions.

Creating a new CONTEXT.md

If no CONTEXT.md exists at the project root, run an elicitation session to discover domain terms:

  1. Explore the codebase to identify domain-specific naming patterns—types, modules, key abstractions, recurring nouns in variable names and comments.
  2. Present findings to the user: "I found these domain concepts in the code: [list]. Which are worth defining? Are any named inconsistently?"
  3. For each confirmed term, resolve four fields: the canonical name, a one-sentence definition, aliases to avoid, and relationships to other terms. These four fields are the minimum needed for an agent to use the term correctly and consistently—fewer leads to ambiguity, more adds noise.
  4. Ask about ambiguities: "Is [term] the same as [similar term], or are they distinct?" Surface these early because unresolved synonyms are the primary source of naming inconsistency in agent-generated code.
  5. Write CONTEXT.md following the format in references/context-format.md.

Updating an existing CONTEXT.md

If CONTEXT.md already exists, read it first. Then:

  1. Identify the new concept or ambiguity that triggered this invocation.
  2. Check whether it fits an existing term (update the definition) or is genuinely new (add it).
  3. If the new concept conflicts with an existing term, add it to "Flagged ambiguities" with a clear resolution.
  4. Preserve all existing content—append, don't rewrite. Rewriting destroys cross-session term provenance and risks silently dropping terms other sessions depend on.

Consuming CONTEXT.md during other work

When exploring or modifying code in a project that has a CONTEXT.md:

  • Read CONTEXT.md before starting work. Use its terms exactly—not synonyms the glossary explicitly avoids. Consistent terminology reduces the cognitive translation cost between domain conversations and implementation.
  • Note any concepts needed for the current task that are missing from the glossary. Flag them as candidates for the next /glossary session.
  • Name variables, functions, types, and files using the glossary's vocabulary.