muki-ai-plugins
A Claude Code plugin set: Figma visual regression, test and code review, project lore and session analysis as an HTML report
- 73 ★ on the repository
- Last commit
Extended access
Why this level
- The session-review plugin reads all local sessions of a project in full, and they can hold code fragments and data
- figma-visual-reviewer calls the Figma API with your token and opens pages in Playwright
Install
- Works in
Manual install
/plugin marketplace add mukiwu/muki-ai-pluginsThen /plugin install with the plugin name, for example review-tests.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The session-review plugin reads all local sessions of a project in full, and they can hold code fragments and data
- figma-visual-reviewer calls the Figma API with your token and opens pages in Playwright
This is third-party code. Review the repository files before installing.
What it does
A set of eight plugins added with one command, `/plugin marketplace add mukiwu/muki-ai-plugins`. The `figma-visual-reviewer` plugin compares a Figma design with a live page: Figma API export, a Playwright screenshot, a pixel diff and a model judgment, with a side-by-side HTML report and multi-viewport support. The `/review` command starts an interactive check. `review-tests` reads an existing test file, looks for blind spots in assertions, mocks and structure, and writes one HTML report to `.review-tests/` without touching code. `review-code` reviews the diff since a chosen point along three axes, standards, spec and bugs, in separate sub-agents, and important findings are confirmed by two independent voters, with a third on a split. A re-review looks only at the new diff and stops after three rounds. `lore` keeps the implicit knowledge of a project in `docs/lore/` and has eight skills: init, consult, capture, guard against the diff before commit, maintain, health check, eval and terminology alignment. `session-review` reads all Claude Code and Codex sessions of a project and builds an interactive report with a skill ranking, diagrams of real workflows and drafts of new skills. For `jev-search` the code lives in a separate repository, and `hyday-pet` and `world-news` are Claude Code mods with no skills.
Who it is for. Developers and QA engineers on Claude Code who need design-to-page comparison, independent test and code review and analysis of their own AI sessions.
Good fit when
- Markup drifted from the Figma design and you need an HTML report with a pixel diff across viewports
- Tests pass but you suspect empty assertions and weak mocks and want the file reviewed
- Before merging a branch you want the diff checked for standards, spec and bugs with confirmation of important findings
- Domain rules and the reasons behind decisions are not written down and should live next to the code
- You want to learn from your sessions which repeated actions deserve to become skills
Not a fit when
- You need a security scanner: the code review looks at standards, spec and bugs
- You need automatic fixes for the problems found: the review plugins only read and diagnose
- You need a multi-agent run outside Claude Code: the `lore` hook and reminder extras are not available there
Example request
Check the tests/cart.test.ts file with the review-tests plugin and show the blind spots in the report.Limitations
The main path targets Claude Code. For Codex, Cursor and Gemini CLI the `npx skills add mukiwu/muki-ai-plugins` command installs only the skill folders: the session-start reminder and pre-commit check of `lore` are not included. The `hyday-pet` and `world-news` plugins are Claude Code mods with no skills, and `jev-search` lives in a separate repository.
No license is stated in the repository. `figma-visual-reviewer` needs Figma API access and Playwright, and `session-review` reads all sessions of a project in full. Function hooks in `jev-search` are early access and need `CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1`. The last commit is from October 20, 2026.
How to disable. Remove a plugin with `/plugin uninstall` and its name, or the whole marketplace. For skills installed through `npx skills add`, delete their folders from the agent skills directory. The `.review-tests/` and `docs/lore/` folders stay in projects and are removed by hand.
FAQ
How do I add the muki-ai-plugins marketplace to Claude Code?
Run `/plugin marketplace add mukiwu/muki-ai-plugins`, then install the plugins you need, for example `/plugin install review-tests`.
How do I install it in Codex, Cursor or Gemini CLI?
With `npx skills add mukiwu/muki-ai-plugins`: the installer asks which skills and which agents to install for. In Codex a skill is invoked as `$name`, for example `$session-review`.
Does review-tests change my code?
No. The plugin only reads the test file and writes an HTML report to `.review-tests/`. Fixes stay with you, and the author recommends driving them through TDD.
How does review-code confirm findings?
Two fresh sub-agents vote on each important finding, and a third is added on a split. Findings are labelled Important, Nit or Pre-existing.
What does session-review show?
A skill ranking by who invoked each, your real workflow as diagrams, quotes on how you build and debug with AI and skill drafts for repeated routines. The report is built from Claude Code and Codex sessions.
Discussion
No questions or reviews yet.
Related
Pluginsshow-your-work
claude-show-your-work
A plugin that shows images and video in the Claude Code terminal, records Playwright checks and compares before and after
The evidence-code-review skill for Codex and Claude Code: evidence-based change review with MUST, SHOULD, BETTER and NITS levels
/pr-review sends a pull request through five independent reviewers, cross-checks findings and builds one report.
A Claude Code plugin that makes an agent check the code around a change and prove the work with a test and independent review
Two skills for Codex and Claude Code: a read-only repository audit and an experiment loop with saved baselines.
A six-skill accessibility plugin: page review, WCAG 2.2 audit with Playwright, audit planning and an improvement roadmap.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| muki-ai-plugins (this card) | no | 73 | ||
| figma-ui-mcp | no | not required | 265 | reads and edits Figma designs over MCP, while this plugin compares a design with a live page pixel by pixel |
| Figma MCP Server | yes | OAuth | 2,179 | gives the agent Figma data, while this adds a visual regression report and test and code review |
README in short
The description lists eight plugins with a task table and a detailed section for each. It then covers installing the marketplace in Claude Code and through `npx skills add` for other agents, noting that extras such as lore hooks do not carry over.