kaizen-review

kaizen-review is a command for Claude Code from littlebearapps/untether. It costs 90 tokens per session (1,436 once invoked), scanned A, original, MIT.

A weekly review process for turning collected improvement notes into proposed tests, documentation, rules, or GitHub issues.

In plain words
What is it for?
Use it to review unaddressed kaizen notes—small process-improvement observations—then prepare proposals for approval.
Why use it?
It helps people sort and combine repeated observations while keeping a human in control of changes to project rules and code.

Command for Claude Code

Written for Claude Code: allowed-tools in frontmatter. Also seen: mentions CLAUDE.md; names the AskUserQuestion tool.

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.

agentmods
npx agentmods add commands/littlebearapps/untether/kaizen-review
Clone the repo
git clone --depth 1 https://github.com/littlebearapps/untether

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 kaizen-review

README.md
[![agentmods](https://agentmods.dev/badge/commands/littlebearapps/untether/kaizen-review.svg)](https://agentmods.dev/commands/littlebearapps/untether/kaizen-review)
Your own site
<a href="https://agentmods.dev/commands/littlebearapps/untether/kaizen-review"><img src="https://agentmods.dev/badge/commands/littlebearapps/untether/kaizen-review.svg" alt="Measured on agentmods" height="20"></a>
Per session 90 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,436 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00090 $0.01436
Opus 5 $0.00045 $0.00718
Sonnet 5 $0.00018 $0.00287
Haiku 4.5 $0.00009 $0.00144

Measured 6d ago against content hash bef366de2112, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

kaizen-review 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 6d 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.

.claude/commands/kaizen-review.md · 117 lines

How it starts

The opening of the file, as written. The whole thing — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are handling /kaizen-review — the weekly promotion pass over the process learnings that /kaizen captured. It is propose-only and human-gated: it recommends, drafts, and files proposals, but it never applies a change to any authoritative surface itself.

User input: $ARGUMENTS

Untether adaptations (read first)

Load docs/kaizen/README.md (the full policy: gate, hierarchy, scales, approval-boundary table) and .claude/rules/kaizen.md. Also .claude/rules/workflow-commands.md. Key points:

  • Authority: propose-only. NEVER auto-edit .claude/rules/, hooks.json, CLAUDE.md, or code. The most it does to an authoritative file is draft a proposed diff into a non-authoritative location (incoming/kaizen-runs/, gitignored) and open a GH issue. Applying the draft is a later human//implement step.
  • Two writes it MAY do: (1) strike a source bullet by editing the collector comment (add ~~…~~ + a → #NN / → dismissed: reason marker); (2) open a GH issue for an accepted promotion. Both are confirm-gated under Untether-mode (state the recommended decision in text; AskUserQuestion returns empty).
  • Idempotent. Re-running must not re-promote an already-struck bullet or double-file an issue. Resolve the collector by exact title, --limit 200.

The promotion gate

Promote a bullet (or a cluster) only if it clears the gate:

Recurrence ≥ 2 OR Severity ≥ S3 OR trivial-high-leverage OR explicit operator request.

Everything else defers (leave the bullet) or dismisses (strike with a reason).

The promotion hierarchy (cheapest sufficient artefact)

Pick the lowest rung that would actually prevent recurrence:

  1. pytest — a regression/guard test (best: makes the learning executable).
  2. doc — a line in an existing doc / reference / FAQ.
  3. rule-draft — a proposed edit to a .claude/rules/*.md (drafted, never applied).
  4. GH issue — when the fix is real work, file it (labelled, severity-tagged).

Read the full file on GitHub · 117 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. 6d ago First seen · 117 lines · 90 tokens per session scan A bef366de2112

Subscribe to this mod's changes

kaizen-review is a command published in the GitHub repository littlebearapps/untether (66 stars, last pushed 5d ago), licensed MIT. It adds 90 tokens to every session and 1,436 once invoked, about $0.0005 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.