Find the term. Keep every requirement.
Precise Terms is an explicitly invoked agent skill that turns verbose descriptions into concise, canonical technical language. It is designed for prompt refinement and terminology discovery, especially across coding, UI, motion design, UX, and architecture.
The skill prioritizes:
- Preserving meaning and explicit requirements
- Selecting the correct established technical terms
- Removing unnecessary wording
The repository root is an Agent Plugins 1.0 package. plugin.json is the portable manifest, while .codex-plugin/plugin.json is a generated compatibility bridge for Codex releases before root-manifest support. Both manifests describe the same explicitly invoked skill.
Recommended Codex installation:
codex plugin marketplace add andydrewie/codex-skills
codex plugin add precise-terms@andydrewie-codex-skillsFor the one-release selective-install compatibility path, use the open skills CLI:
npx skills add andydrewie/precise-terms -g -a codexInvoke the skill explicitly:
Use $precise-terms to rewrite these instructions with concise, canonical technical terminology:
[instructions]
For terminology discovery:
Use $precise-terms to identify the canonical technical term for:
[description]
Verbose:
Start by choosing one small part of the feature and making that part work through the interface, backend, and database.
Refined:
Start with a thin vertical slice spanning the UI, backend, and persistence layer.
- Explicit invocation only; ordinary editing or clarification requests do not activate it.
- Preserve every stated requirement, constraint, qualifier, number, exception, and acceptance criterion.
- Do not invent jargon or infer a named technique from a vague desired outcome.
- Do not execute refined instructions unless the user separately asks for execution.
The normative behavior is defined in SKILL.md.
Run the dependency-free package validator before publishing changes:
python3 scripts/validate_package.pyMIT