claude

guardrails

The authority layer for technical delivery: previews before destructive or database-writing commands (Bash and PowerShell), a hard scope gate on database file edits only, and a standing block on git commit/push/PR — the agent drafts commit messages and PR descriptions as text, with no AI attribution. Ships SessionStart, UserPromptSubmit, PreToolUse and PostToolUse hooks.

0 on GitHubby Kaizen Analytixv0.9.0updated 1 month agotechnical-delivery
/

Plugin guardrails from marketplace kaizen-plugins in kaizenanalytix/kaizen-claude-plugins. Commands read from the repo's manifest — 1 month ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (5)

  • failure-triage

    What to do when a test, build, lint, or command fails — report the real output, sort the failure into test-was-wrong / real-defect / genuinely-ambiguous, state the explanation explicitly as a hypothesis rather than a finding, and ask before changing code. Covers the hypothesis report format, the two

  • guardrails

    Reference skill defining the mandatory guardrails every skill in this marketplace follows — what the agent may do on its own authority and what it must put in front of the user first. Seven ID'd families: consent (C), preview-before-permission (P), data and destructive operations (D), verification a

  • permission-preview

    Shows the user what a command will actually do — and what approving it will actually authorise — before the permission prompt appears. Covers the six-field preview block, running a read-only probe first to get a real sample instead of an invented one, resolving what a connection string or context ac

  • question-protocol

    How the agent asks the user things — batched multiple-choice rather than open-ended, a recommended default on every option, understanding restated before the question so a misread premise gets caught first, an explicit statement whenever a default is taken on the user's behalf, and re-asking rather

  • subtask-orchestration

    Splits a large task into stages and subtasks small enough for one context, each dispatched with acceptance criteria declared up front, verified against those criteria rather than against how good the summary reads, and memoised so a retry reuses finished work instead of re-solving it. Also governs b

Hooks (6)

  • SessionStart

    node "${CLAUDE_PLUGIN_ROOT}/scripts/guardrail-contract.js"

  • UserPromptSubmit

    node "${CLAUDE_PLUGIN_ROOT}/scripts/scope-first.js"

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/scripts/preview-gate.js"

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/scripts/write-gate.js"

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/scripts/agent-dispatch-gate.js"

  • PostToolUse

    node "${CLAUDE_PLUGIN_ROOT}/scripts/failure-triage.js"

View source on GitHub

More from kaizenanalytix/kaizen-claude-plugins

All 18 plugins →

frontend

claude

Domain-modular React apps (Vite, React Router, RTK Query, shadcn/ui, Vitest), with thin architecture adapters for Vue and Angular.

18 skills

0

backend

claude

Clean-architecture FastAPI backends with testing, bootstrap and best practices, plus thin adapters for Node.js, NestJS, Django and Spring.

13 skills

0

Stack-agnostic architecture principles (clean-architecture zones, state, test pyramid), the greenfield kickoff orchestrator, and the adopt-vs-defer decision for existing codebases. Ships SessionStart and PreToolUse hooks.

7 skills · hooks

0

Analyze & Design phase: business requirements, technical design docs, data models, solution-overview, data-flow, API-sequence and architecture diagrams, plus a build handoff package.

8 skills

0

Kvantum Element X pre-load prep: raw client extracts plus template in, validated load-ready file and an input-review dashboard out.

6 skills

0

Chooses how an app should ship before writing anything: Dockerfiles, CI/CD pipelines, dev/staging/prod config, secrets, and rollback.

5 skills

0

Similar plugins

Commands for git commit workflows including commit, push, and PR creation

3 commands

anthropics/claude-code 155,476

Professional UI/UX design intelligence for AI coding assistants. Includes searchable databases of styles, colors, typography, charts, and UX guidelines for React, Next.js, Astro, Vue, Nuxt.js, Nuxt UI, Svelte, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui, Jetpack Compose, Angular, Laravel, JavaFX, and Three.js.

7 skills

nextlevelbuilder/ui-ux-pro-max-skill 148,759

storybook

claudecodex

Build, preview, and test UI components with Storybook.

7 skills

storybookjs/storybook 91,552

Self-learning vector database — HNSW, FlashAttention-3, Graph RAG, hybrid search, DiskANN, and Brain AGI

4 skills · 1 command · 1 agent

ruvnet/ruflo 81,589

diagram-design

claudecodex

Editorial diagrams in your brand, as self-contained HTML and SVG. Architecture, flowcharts, sequence, state machines, ER and database schemas, timelines, swimlanes, org charts, Gantt, Wardley maps, user journeys, service blueprints, UML class, lifecycle phase maps, and charts from bar, line and scatter to Sankey, treemap and heatmap, plus many more. Redraws draw.io, Mermaid, Excalidraw and PlantUML files. Exports SVG and PNG.

1 skill · 7 commands

cathrynlavery/diagram-design 66,757

Perform AI code review on Git diffs — supports workspace changes, branch ranges, and single commits with concurrent per-file analysis, codebase search, and deep context-aware review.

2 commands

alibaba/open-code-review 50,068

Maintain guardrails?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins18 plugins