/wbCheck: Practical Guide 🛠️

The problem /wbCheck solves

You have 23 /wb* commands and multiple AI models. Some models are cheap but may not understand your codebase. If you send /wbRefactor to a model that doesn't know Vue 2.7 from Vue 3, it will rewrite your Options API as Composition API and break everything.

The fix: A 30-second quiz that catches this before damage happens.

The security model

YOU: Have the answer key ([private quiz tool])
WORKER: Only has access to workspace files (source code)
RULE: Worker must prove understanding from CODE, not from a cheat sheet

The worker never sees the answers. It can only pass by actually reading WBC.js, package.json, TEST_REPORT.md, etc. If it fabricates answers, you catch it because you have the keywords to compare against.

Real example (generic — no answer-key data shown)

Before running /wbTest with the agent V4 Flash:

$ wbcheck wb-core wbTest

# Script outputs 6 random questions (3 identity + 3 testing)
# You copy them to the agent
# the agent reads the source code and answers
# You press Enter → compare against your private answer key

How you grade:

• If the model gives exact numbers matching TEST_REPORT.md → ✅ (it read the file)
• If the model gives round numbers or wrong counts → ❌ (hallucinating from training data)

Command-category mapping

The script knows which topics matter for each command:

Command	What it asks about	Why
/wbTest	identity + testing	Must know test runner, counts, failing files
/wbRefactor	identity + architecture	Must know mixins, this-binding, Vue version
/wbAudit	identity + architecture + security	Must know everything
/wbSecure	identity + security	Must know safeEval, DOMPurify, tier bypass
/wbRelease	identity + build	Must know dist folders, npm name, Vite

When to skip

Don't quiz a model you've already verified on this package with no code changes since. Don't quiz the agent/Antigravity (it has persistent context). Only quiz new or cheap models on unfamiliar packages.

---