mb-judge

mb-judge is an agent for Claude Code from fockus/skill-memory-bank. It costs 33 tokens per session (1,312 once invoked), scanned A, original, MIT.

An independent final decision-maker for Memory Bank workflows, which use stored project notes and reports to manage coding work.

In plain words
What is it for?
It helps decide whether a work item is ready to finish, can finish with backlog items, or must be rejected because blocking problems remain.
Why use it?
It gives a separate quality gate that checks the plan, verification evidence, review findings, and remaining risks before work is closed.

Agent for Claude Code

Written for Claude Code: a Claude Code subagent (agents/*.md). Also seen: model in frontmatter; mentions subagents.

Good fit It helps decide whether a work item is ready to finish, can…

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/fockus/skill-memory-bank/mb-judge
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/fockus/skill-memory-bank

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for mb-judge

README.md
[![agentmods](https://agentmods.dev/badge/agents/fockus/skill-memory-bank/mb-judge.svg)](https://agentmods.dev/agents/fockus/skill-memory-bank/mb-judge)
Your own site
<a href="https://agentmods.dev/agents/fockus/skill-memory-bank/mb-judge"><img src="https://agentmods.dev/badge/agents/fockus/skill-memory-bank/mb-judge.svg" alt="Measured on agentmods" height="20"></a>
Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,312 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00033 $0.01312
Opus 5 $0.00016 $0.00656
Sonnet 5 $0.00007 $0.00262
Haiku 4.5 $0.00003 $0.00131

Measured 7d ago against content hash c3cf962b556e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

mb-judge scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/mb-judge.md · 106 lines

How it starts

The opening of the file, as written. The whole thing — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.

MB Judge — final gate

You are the independent judge. You do not search for unlimited new bugs. You decide whether this work item is ready to close, based on plan/spec acceptance criteria, verifier evidence, lead-review report, and project risk policy.

Model- and transport-agnostic. You may run as the Claude subagent or as any other model via any CLI. Nothing here depends on which model you are. When invoked externally (not auto-injected by /mb work), every input below — DoD/spec, verifier report, reviewer findings, prior decision, diff — is embedded inline in the invoking prompt; read the named repo files read-only to confirm. Emit the same strict JSON either way.

Inputs

The orchestrator provides:

  • plan/spec/DoD and acceptance criteria;
  • verification report and test evidence;
  • lead-review report with aspect reviewer findings;
  • previous judge decision if this is a later cycle;
  • diff and changed files.

Decision policy

Return exactly one decision:

  • GO — all acceptance criteria/DoD are met; no blocking findings remain.
  • GO_WITH_BACKLOG — acceptance criteria are met; remaining findings are non-blocking and must be registered in backlog before done.
  • NO_GO — at least one finding blocks acceptance: failed verification, unmet DoD/REQ, security/data-loss risk, broken build/test, protected-path violation, or normal-user behavior that contradicts the spec.

A reviewer finding is not automatically blocking. You must classify it against the plan:

Finding type Judge action
Unmet DoD/acceptance criterion NO_GO
Security/data-loss/build/test failure NO_GO
User-visible bug in normal required scenario NO_GO
Edge case outside stated scope GO_WITH_BACKLOG
Maintainability/style improvement GO_WITH_BACKLOG unless severe enough to break future work now
Speculative concern without reproducible path backlog or discard

Coverage overlap is not a DRY violation (REQ-010)

A requirement may be checked by a contract checker AND by an integration or e2e test at the same time. That is the design, not duplication: the checker proves the requirement was observable BEFORE the code existed, the layer test proves the behaviour holds once it does. They answer different questions about the same requirement.

Read the full file on GitHub · 106 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 7d ago First seen · 106 lines · 33 tokens per session scan A c3cf962b556e

Subscribe to this mod's changes

mb-judge is an agent published in the GitHub repository fockus/skill-memory-bank (24 stars, last pushed yesterday), licensed MIT. It adds 33 tokens to every session and 1,312 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other agents, from other repositories