mb-review
Original:🇺🇸 English
Translated
1 scriptsChecked / no sensitive code detected
Review a Memory Bank with fresh-context specialists and produce a prioritized fix list.
12installs
Sourcemrvladd-d/memobank
Added on
NPX Install
npx skill4agent add mrvladd-d/memobank mb-reviewTags
Translated version includes tags in frontmatterSKILL.md Content
View Translation Comparison →mb-review — Multi-expert Memory Bank review
- What it does: runs independent reviewers over architecture, scope, backlog, security, and Memory Bank discipline.
- Use it when: you need a clean review gate before execution, after bootstrap, or before trusting the docs.
- Input: an existing .
.memory-bank/ - Output: reviewer reports in plus a synthesized list of fixes and review verdicts.
.tasks/TASK-MB-REVIEW/
Goal
Detect gaps, contradictions, broken traceability, and non-compliance early.
Preconditions
- exists.
.memory-bank/
Process
1) Create review task folder
Create:
.tasks/TASK-MB-REVIEW/
2) Spawn reviewers (fresh contexts)
Spawn these subagents in parallel (max 5–7):
- Architect —
./agents/shared-review-architect.md - Scope/RTM —
./agents/shared-review-scope.md - Plan/backlog —
./agents/shared-review-plan.md - Security —
./agents/shared-review-security.md - MBB compliance —
./agents/shared-mb-reviewer.md - Code quality (conditional: if repo has code) —
./agents/shared-review-code.md
Each reviewer must:
- write a detailed report to
.tasks/TASK-MB-REVIEW/ - return only a short summary + verdict to the orchestrator
3) Synthesize and decide
As orchestrator:
- combine findings
- deduplicate
- rank issues P0–P3
- produce a concrete fix plan
If the repo is preparing for :
/autonomous- treat unresolved P0/P1 issues as blocking
- do not allow batch execution until the final verdict is
APPROVE
4) Gate
If any reviewer returns :
REJECT- fix MB
- re-run mb-review
Definition of done
- contains the reviewer reports.
.tasks/TASK-MB-REVIEW/ - Orchestrator produced an actionable prioritized fix list.
- Final verdict: APPROVE.