| name | ste-dictionary-generate |
| type | Skill |
| title | Extensible STE Dictionary Generator |
| description | Automatically generate a project-specific Extensible STE Dictionary (.ste-dictionary.yaml) from a codebase or document set. Cross-references the `information-architecture` skill suite (`vocabulary-overlap-analysis`, `thesaurus-generate`, `named-entity-normalization`) to discover proprietary technical terms, software/hardware entities, UI names, and action verbs, formatting them into ASD-STE100 schema. Use when initializing STE documentation for a new repository or updating an existing project dictionary. |
| tags | ["ste","asd-ste100","dictionary-generator","controlled-vocabulary","information-architecture","taxonomy"] |
| status | verified |
| generated | 2026-08-06T00:00:00.000Z |
| sources | ["https://www.asd-ste100.org/","https://github.com/dandye/information-architecture"] |
| resource | https://github.com/dandye/ste-writing-style |
| personas | ["information-architect","taxonomist","technical-writer","software-engineer"] |
STE Dictionary Generation Skill
This skill automates the creation and updating of an Extensible STE Project Dictionary (.ste-dictionary.yaml or ste-dictionary.json) by extracting domain terminology from project source files or documentation corpora.
It leverages the information-architecture skill suite to analyze text, isolate unique domain entities, and structure them into approved Technical Names (TN) and Technical Verbs (TV).
Inputs
TARGET_PATH: The directory or file path of the project codebase or documentation corpus to analyze (e.g. docs/ or src/).OUTPUT_FILE: (Optional) Target dictionary file path (default: .ste-dictionary.yaml in repo root).REFERENCE_PATH: (Optional) Secondary document path to contrast against for vocabulary overlap analysis.
Workflow
Step 1: Vocabulary Extraction via information-architecture
-
Execute the vocabulary-overlap-analysis skill from information-architecture (/ia:vocab-overlap):
- Pass
TARGET_PATH to extract unigrams, bigrams, acronyms, and proper nouns.
- Filter out standard English stopwords and baseline web jargon to isolate proprietary product names, software terms, and specialized domain processes.
-
Execute thesaurus-generate (/ia:thesaurus) or named-entity-normalization (/ia:entity-normalize):
- Group spelling variants, acronyms, and alternate phrasings under a single Preferred Term.
- Map non-preferred variants as Forbidden Synonyms.
Step 2: Categorize Terms into STE Schema
Classify all extracted preferred terms into ASD-STE100 categories:
-
Technical Names (TN):
Assign POS noun and assign one of the allowed STE categories:
software: e.g. kernel, payload, repository, pod, .
