# Bdc Assessment Designer

> Design and safely create complete, grounded LMS quizzes without using the platform LLM gateway.

- Skill: `big-data-club/bdc-assessment-designer` (Agent Skill)
- Install (CLI): `npx skillmds@latest add big-data-club/bdc-assessment-designer`
- Raw SKILL.md: https://api.skillmd.com/api/skills/big-data-club/bdc-assessment-designer/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- License: CC-BY-NC-SA-4.0 OR LicenseRef-BDC-Commercial
- Author: Big-Data-Club (https://skillmd.com/u/big-data-club)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/big-data-club/bdc-assessment-designer

---


# BDC assessment designer

1. Resolve an owned course with `list_my_courses`, then resolve nodes with `list_knowledge_nodes`.
2. Search the newly posted lessons with `search_course_materials` before writing. Record precise `source_refs` such as `content:2559#kv-cache` on questions that depend on lesson facts; never invent facts absent from those sources.
3. For a standard complete quiz, author exactly 10 questions in this order: 1–4 `remember`/easy, 5–6 `understand`/medium, 7 `apply`/medium, and 8–10 `analyze`/hard calculation. Questions 8–10 must each include at least one lesson `source_ref`.
4. Every question needs 2–6 distinct options, one valid zero-based `correct_index`, a useful explanation (including easy questions), and points. Do not force all correct answers into option A; the tool also balances saved answer positions defensively.
5. Avoid trick wording, sensitive profiling, and answer clues. Verify equations, units, rounding rules, and numerical distractors against the lesson evidence.
6. Before writing, show one preview containing title, target course/section, question count, schedule, duration, maximum attempts, passing score, shuffle settings, and whether it will publish. Ask for explicit confirmation.
7. After approval, call `mcp_batch_generate_quiz` once with `quiz_settings`. Supply `section_id`, `time_limit_minutes`, `max_attempts`, optional timezone-aware `available_from`/`available_until`, shuffle flags, and `publish`. The tool saves bank items and assembles the complete quiz; do not tell the teacher to manually assemble it.
8. Use `publish: false` when review is still required. Use `publish: true` only when the teacher explicitly approved publication. Report returned `quiz_id`, `content_id`, state, and preserved `bank_item_ids`; never claim publication unless the returned state is `PUBLISHED`.

Example settings:

```json
{
  "section_id": 378,
  "time_limit_minutes": 30,
  "max_attempts": 10,
  "passing_score": 80,
  "available_from": "2026-09-02T08:00:00+07:00",
  "available_until": "2026-09-30T23:59:00+07:00",
  "shuffle_questions": false,
  "shuffle_answers": true,
  "publish": false,
  "enforce_bloom_ladder": true,
  "require_advanced_source_refs": true
}
```