Skip to content

/wbReview — Live Demo () ​

What /wbReview would actually surface on wb-labs right now. The uncommitted state of the workspace as of the current session is large — a docs synchronization run with dozens of new files in frontEnd/wbc-ui/core2/apps/wb-flow/flow.wbc-ui.com/. That uncommitted state is the most realistic review target available.


1. Live target ​

FieldLive value
Uncommitted stateMany new files under frontEnd/wbc-ui/core2/apps/wb-flow/flow.wbc-ui.com/ (the v4 generation in progress) plus a few modified the agent files
Active planreports/<date>/plans/plan_wb-core_20260504.md (3 rows, all about wb-core code work)
Closest "plan of record" for this sessionfrontEnd/wbc-ui/core2/apps/wb-flow/wb-flow-docs/claude_transmission_sync_v4.md (the transmission spec — used as a quasi-plan)
What memory says about docs scopeproject_docs_edition.md records the structural rules for docs/

There's an interesting tension to surface: the active plan is about wb-core code, but the uncommitted work is documentation generation. A naive review would flag every new docs file as "scope creep" against the wb-core plan. A smart review reads the transmission spec and recognizes the docs work as a separate, well-defined session.


2. What each target form would resolve to today ​

TargetLive resolution
/wbReview (no arg)Reviews ~17+ uncommitted .md files. Maps them against the transmission spec, not the wb-core plan.
/wbReview --planStrict mode against the wb-core plan → would fail every docs file as scope creep. Wrong tool for this session.
/wbReview frontEnd/wbc-ui/core2/apps/wb-flow/flow.wbc-ui.com/Scoped review of just the new docs. Realistic.
/wbReview "the docs sync"Refused — free-text. Suggests the path.
/wbReview HEAD~1Reviews the last commit (e07bb93 chore(core2): root cleanup, context integration, and daily reporting).

The first two cells are the most instructive. The --plan flag is the wrong choice for this session — the active plan is about a different scope. The agent surfaces this rather than mechanically running strict mode against a mismatched plan.


3. Per-flag behavior, applied live ​

Flag combinationLive result
/wbReview (advisory)Reviews docs files, treats claude_transmission_sync_v4.md as the de-facto plan, flags any divergences from its 6 verification-checklist groups.
/wbReview --plan against plan_wb-core_20260504.mdHalts with "this plan does not cover the current work — review this session against the transmission spec instead."
/wbReview frontEnd/wbc-ui/core2/apps/wb-flow/flow.wbc-ui.com/ --actRanked action file: which generated docs are weakest and most need revision.
/wbReview frontEnd/wbc-ui/core2/apps/wb-flow/flow.wbc-ui.com/ --wbPlanAdds plan rows for revision work. Risky here — would flood the wb-core plan with docs-revision work. The agent surfaces this as a warning.

That last cell is the interesting one. --wbPlan mutates the active plan, which currently covers wb-core. Adding 17 docs-revision rows to a 3-row code plan is a memory-aware misuse. The agent will produce a refuse-with-suggestion: "The active plan is wb-core code work. Run /wbPlan --focus="docs" first to create a docs-specific plan, then re-run /wbReview --wbPlan."


4. Pipelines ​

📋/wbReview Live Execution Simulator
Advisory review of the docs sync in progress/wbReview frontEnd/wbc-ui/core2/apps/wb-flow/flow.wbc-ui.com/
> /wbReview frontEnd/wbc-ui/core2/apps/wb-flow/flow.wbc-ui.com/
[SYSTEM] Target: frontEnd/wbc-ui/core2/apps/wb-flow/flow.wbc-ui.com/ (uncommitted changes).
[CONTEXT] No matching active plan (plan_wb-core_20260504.md is for code).
Falling back to: claude_transmission_sync_v4.md as quasi-plan.
[MAP] Touched files → spec sections:
commands/wbWork/{exhaustive_simulation,live_demo}.md → Part 2
commands/wbValid/{...}.md → Part 2
commands/wbExplain/{...}.md → Part 2
commands/wbAudit/{...}.md → Part 2
commands/wbDebug/{...}.md → Part 2
commands/wbTest/{...}.md → Part 2
commands/wbReview/{...}.md → Part 2 (THIS FILE)
commands/wbGit/wbGit_practical.md (modified) → Part 1
commands/wbGit/wbGit_examples.md (modified) → Part 1
concepts/flags_and_shortcuts.md (modified) → Part 1
## Findings (advisory)
### 1 — Part 1 done, Part 2 in progress
**Severity:** Informational
**Status vs spec:** Part 1 (wbGit + flags) — complete. Part 2
(dual-file v4) — 7 of 30 commands so far (Execution & QA group
mid-run).
**Comment:** Progress matches spec ordering (QA group declared first
in section 3 verification checklist). No drift.
### 2 — wbGit dual files not yet generated
**Severity:** P2
**Files missing:** commands/wbGit/wbGit_exhaustive_simulation.md
commands/wbGit/wbGit_live_demo.md
**Why:** Spec Part 2 verification checklist lists wbGit in the
Execution & QA group. The two updated wbGit files (practical,
examples) are Part 1 artifacts, not Part 2.
**Recommendation:** Don't skip these. They belong in this batch.
### 3 — flags_and_shortcuts.md edits look right
**Severity:** ✅ pass
**Comment:** Rule 5 added (`-L` mnemonic). wbGit row updated to
10 flags. Matches spec Part 1 section 3.
[OK] 3 findings, advisory. 1 informational, 1 P2 actionable, 1 pass.
[OK] No mutation. Mid-session review — keep going.
⚠️ Note: Running `/wbReview` against the live uncommitted state right now:

💠 Pipeline Advisory review of the docs sync in progress ​

Running /wbReview against the live uncommitted state right now:

💠 Pipeline The misuse the agent should refuse ​

A user, mid-session, runs /wbReview --plan thinking it'll gate the docs work:

💠 Pipeline Reviewing the last committed change ​


5. What would refuse today ​

TriggerLive response
/wbReview with no uncommitted changes and no commit refHalt. ❌ Nothing to review.
/wbReview "the docs work"Refused with disambiguation; suggests the path.
/wbReview --plan against the wrong planRefuses with explanation; suggests advisory mode or making a matching plan first.
/wbReview --plan --wbPlan mutating into a mismatched planRefuses with same logic; suggests /wbPlan --focus=... first.
/wbReview on a diff that's only generated filesOne-line notice. Skips generated files.
/wbReview on the docs corpus, repeatedly within a sessionEach run re-evaluates against current uncommitted state. No state cached.

The pattern: /wbReview is a plan-aware critic and refuses to run strict mode against the wrong plan. It reads the active plan, compares to the transmission spec when one exists, and surfaces drift advisorily by default. When --plan is misused (mismatched scope), it refuses with a constructive next step. The chaining flags (--act, --wbPlan) become useful only after the right plan is in place — until then, the advisory output is the negotiation surface.