wb-flow

Appearance

/wbAudit — Command Hub ​

/wbAudit is the WB-Labs honesty engine. It reads actual source files — not summaries, not plans — and returns a scored, ranked assessment of whether code is ready to ship. It is the command you run before /wbRelease, not after something breaks.

Detecting and correcting drift

/wbAudit finds the gap between what a scope claims and what it does — then the plan it emits is what closes it.

🎯 Strategic Position ​

Every other command in the suite tells you what to do. /wbAudit tells you where you actually are. That distinction matters because the default failure mode of an AI reviewer is agreeableness — "looks good overall, minor improvements possible" — which is always wrong for a non-trivial package. Every active package has at least one MAJOR finding.

  • Before a release — a folder audit on the package being shipped.
  • After inheriting code — a baseline of the debt you just took on.
  • Before a refactor — a file audit on the file you are about to change.

🛠️ Operating Modes ​

ModeTriggerOutput
Strategic/wbAudit <folder>Score /10, findings ranked BLOCKER / MAJOR / MINOR, ship recommendation
Surgical/wbAudit <file>Flaws in one file, plus an existing-vs-suggested comparison table
Self-correct/wbAudit <previous_audit.md>Verifies and repairs the prior report in place

✅ What a useful audit contains ​

A useless audit is missing at least one of these five:

  1. A score with a ship / don't-ship recommendation.
  2. Findings ranked by severity.
  3. Specific file and line references.
  4. A "what this audit did NOT check" section.
  5. The next command to run.

If any is absent, re-run it with "harsher."

🚫 What it cannot do ​

Not thisUse instead
Runtime performance/wbTest --profile
Adversarial security/wbSecure
Whether a feature should exist/wbVision
Why something is broken/wbDebug
Verifying one specific change/wbReview

It also surfaces open architectural decisions rather than resolving them. An audit describes reality; it does not choose your architecture.

📚 Reading Order ​

  1. ELI5 — the one-paragraph mental model.
  2. Practical — the two modes and when each applies.
  3. Expert — why severity ranking beats a single score.
  4. Examples — annotated audit transcripts.
  5. Exhaustive simulation · Live demo.
  • wbAudit.md — the command reference this hub orients you around.
  • /wbActOn — turns an audit into a ranked execution order.
  • /wbPlan — turns audit findings into a task table.
  • /wbNext — ranks what to do once an audit lands.

Quick Reference ​

bash
/wbAudit packages/wb-core/            # strategic: score + ranked findings
/wbAudit packages/wb-core/src/WBC.js  # surgical: one file
/wbAudit <folder> --act               # chain into /wbActOn
/wbAudit <folder> --wbPlan            # chain into plan generation
/wbAudit <folder> --ideas             # route P3 findings to the idea pipeline

← Home · Commands · Install | wb-flow on npm · flow.wbc-ui.com · wi-bg.com