zao-sheng/ontology-playground-zh-dev/name-generator

Name Generator Skill

Goal

Generate person names for this repository from the approved local name fixture:

data/reference/FNF-2026-06-01-01002-0268.csv

Do not invent person names. Every person name used in examples, sample data, quests, tests, demos, docs, or generated ontology content must come from the CSV FullName column.

Source File

CSV columns:

FirstName,LastName,FullName,FirstNameNative,LastNameNative,FullNameNative,Gender,Language

Use FullName by default. Use FullNameNative only when the user explicitly asks for native-script names or locale-specific display text.

Workflow

1. Decide how many names are needed

Identify the role and quantity from the task, for example:

• sample customers
• employees or managers
• patients or clinicians
• students or instructors
• reviewers, approvers, assignees, or contributors

If the task does not specify quantity, use the minimum number needed for the example or test.

2. Read names from the CSV

Use the CSV fixture as the only source. A quick shell-friendly way to inspect the first approved names is:

awk -F, 'NR > 1 { print $3 }' data/reference/FNF-2026-06-01-01002-0268.csv | head

For random sampling:

awk -F, 'NR > 1 { print rand() "\t" $3 }' data/reference/FNF-2026-06-01-01002-0268.csv | sort -n | cut -f2- | head -n 5

If names with commas or quotes are ever added to the CSV, use a proper CSV parser instead of field splitting.

3. Fit names to the scenario

• Choose distinct names for distinct entities.
• Keep the selected names stable within a scenario so queries, expected results,

docs, and sample instances stay consistent.

• Do not alter spellings unless the surrounding file has a strict ASCII-only

convention. If ASCII is required, choose names from the CSV that are already ASCII-compatible.

• Email addresses and IDs may be generic (customer001@example.com) and do not

need to use the person's name.

4. Update all dependent examples

When replacing a name in code or content, update every coupled surface:

• sample instances
• query prompts and curated query matches
• expected test strings
• rendered docs or generated content source files
• catalogue examples or learning materials

Regenerate compiled artifacts when source content changes:

npm run catalogue:build
npm run learn:build

Validation

Before finishing a name-generation or name-replacement task:

1. Verify each selected name appears in the CSV FullName column.
2. Search for removed placeholder names to ensure no stale references remain.
3. Run focused tests for touched code paths.
4. Run npm run build when generated catalogue or learning output changes.

Done Criteria

• [ ] All person names used by the task come from data/reference/FNF-2026-06-01-01002-0268.csv.
• [ ] No invented placeholder names remain in the touched examples.
• [ ] Related prompts, sample data, expected results, and tests are consistent.
• [ ] Relevant tests or build commands have passed, or any skipped validation is clearly reported.