thermo-nuclear-code-quality-review skill: a harsh maintainability review from Cursor

Makes your agent review your branch with a strict maintainability rubric that hunts for dramatic simplifications.

Installs: 22.8K; on skills.sh
Stars: 15K; on GitHub
Updated: Nov 2026; last repo commit

thermo-nuclear-code-quality-review gives your agent a strict code review rubric. It audits the current branch's changes for abstraction quality, giant files and spaghetti conditions, and pushes for "code judo" moves that make the implementation much simpler without changing behavior.

The SKILL.md sets disable-model-invocation: true, so you call it yourself with phrases like "thermo-nuclear code quality review". It bundles no scripts and names no required tools. It is part of Cursor's thermos plugin, where a separate skill covers correctness and security.

Good for

Developers who want a harsh second look at their branch before a PR
Reviewers who care about file size, spaghetti branching and layering
Teams that want review feedback on structure, not nits

Install thermo-nuclear-code-quality-review

Claude Code

Copy this prompt.

Install the thermo-nuclear-code-quality-review skill in Claude Code. Run this command:

npx skills add https://github.com/cursor/plugins --skill thermo-nuclear-code-quality-review

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js.

From its listing on skills.sh. Source.

What's in the skill

When it loads

Run an extremely strict maintainability review for abstraction quality, giant files, and spaghetti-condition growth. Use for a thermo-nuclear code quality review, thermonuclear review, deep code quality audit, or especially harsh maintainability review.

Sections in SKILL.md

Core Prompt
Non-Negotiable Additional Standards
Primary Review Questions
What to Flag Aggressively
Preferred Remedies
Review Tone
Output Expectations
Approval Bar

Key facts

Author: Cursor (thermos plugin)
Repository: cursor/plugins (thermos/skills/thermo-nuclear-code-quality-review)
Installs on skills.sh: 22,793 as of 2026-11-02
GitHub stars: 14,989 for the whole cursor/plugins repo, 2026-11-02
License: MIT per the thermos plugin.json (GitHub detects no license file for the skill)
Last commit to the repo: 2026-11-01
What is inside: 1 file (SKILL.md only), part of the thermos plugin

What it includes

Deep audit of the branch, ambitious about structure
Start from a baseline prompt: rethink the implementation to improve abstractions and modularity without changing behavior, and be thorough.
Read · Core prompt
Code judo, 1k-line rule, no spaghetti, typed boundaries
Rules on simplification, files crossing 1,000 lines, ad-hoc conditionals, direct boring code, type cleanliness, canonical layers, and sequential or non-atomic orchestration.
Read · Eight standards
13 questions asked of every meaningful change
Examples: is there a code judo move, did the diff enlarge a file past a healthy size, is the logic in the right layer.
Read · Primary review questions
What to escalate and what fixes to suggest
Lists patterns to flag aggressively, such as thin wrappers and one-off flags, and preferred remedies such as deleting a layer or splitting a file.
Read · Flags and remedies
Structural issues first, few high-conviction comments
Seven priority levels, and an approval bar that blocks a PR which pushes a file past 1,000 lines or adds tangled branching unless the author justifies it.
Try: “Run a thermo-nuclear code quality review of this branch.”
Read · Output priority and approval bar

From the official docs

Questions

What does thermo-nuclear-code-quality-review do?

It makes your agent review the current branch with a very strict maintainability rubric. The rubric covers abstraction quality, file size, special-case branching, type boundaries and layering, and pushes for simpler designs.

How do I invoke it?

Ask for it by name. The SKILL.md sets disable-model-invocation, which means the agent does not pick it up on its own. Its description lists trigger phrases: "thermo-nuclear code quality review", "thermonuclear review", "deep code quality audit" and "especially harsh maintainability review".

How do I install it in Claude Code, the Claude app or Codex?

Use the install box at the top of this page. The thermos README documents /add-plugin thermos for Cursor agents, which also adds its two other skills and agents. The SKILL.md itself does not name which clients it works in.

Does it change my code?

The SKILL.md describes a review that produces findings and suggested remedies, not edits. Its output section lists finding priorities. How your agent acts on those findings depends on your prompt.

What is the 1,000-line rule?

The skill treats a PR that pushes a file from under 1,000 lines to over 1,000 as a presumptive blocker unless the author gives a strong structural reason.

How is it different from thermo-nuclear-review?

The thermos README describes thermo-nuclear-review as a deep audit for bugs, breakages, security and feature-gate leaks. This skill covers maintainability only. The thermos skill runs both reviews in parallel.

Is it safe to install?

It is one markdown file with no bundled scripts. It tells the agent to review a branch and does not list shell commands. If you ask your agent to apply its suggestions, the resulting refactors can be large, so review them like any big diff.

Information

updated Nov 2, 2026

Repository: cursor/plugins
Folder: thermos/skills/thermo-nuclear-code-quality-review

cursor/plugins skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 3

thermo-nuclear-code-quality-review SKILL.md · github.com
thermos README · github.com
thermos plugin.json · github.com