thermos
cursor/plugins
Run parallel thermo-nuclear code reviews and synthesize findings for comprehensive branch audits.
What is thermos?
Thermos launches two specialized review subagents in parallel—one for bugs, security, and DevEx risks, and one for code quality and maintainability—then synthesizes their findings into a unified report. Use it for comprehensive audits of pull requests or branches when you need both safety and health assessments.
- Launch bug/security and code-quality review subagents in parallel as background tasks
- Gather diffs and file context automatically for reviewer evaluation
- Deduplicate and weight overlapping findings across both reviewers
- Synthesize results into a unified verdict with highest-signal issues first
- Resolve disagreements between reviewers using reasoned judgment
How to install thermos
npx skills add https://github.com/cursor/plugins --skill thermosHow to use thermos
- 1.Provide the review scope: a PR, branch name, or list of changed files
- 2.Let Thermos gather the diff and relevant file excerpts
- 3.Wait for both subagents to complete their background analysis
- 4.Review the synthesized findings, prioritized by signal and deduplicated across reviewers
- 5.Act on the unified verdict and highest-priority issues
Use cases
- Audit a pull request for both security risks and code-quality regressions simultaneously
- Review a feature branch before merge to catch bugs, breakages, and maintainability issues
- Perform a combined DevEx and codebase-health assessment on recent changes
- Detect feature-flag leaks, spaghetti code, and abstraction problems in one pass
- Validate that refactors improve structure without introducing hidden bugs
- Code reviewers and maintainers
- Engineering leads performing branch audits
- Teams requiring both security and quality gates before merge
- Developers seeking comprehensive feedback on their changes
thermos FAQ
The thermo-nuclear-review-subagent focuses on bugs, security, DevEx regressions, and branch-audit risks. The thermo-nuclear-code-quality-review-subagent focuses on maintainability, structure, file-size growth, and codebase-health risks.
Parallel execution saves time by running both reviews concurrently, then synthesizing results once both finish, rather than waiting for one to complete before starting the other.
It weights overlapping findings more heavily, resolves disagreements using reasoned judgment, and surfaces the unified verdict with highest-signal issues first.
No. Thermos gathers the diff and file context automatically based on the review scope you provide (PR, branch, or changed files).
If individual summaries are already visible, Thermos avoids restating them wholesale and instead surfaces the unified verdict and remaining high-signal findings.
Full instructions (SKILL.md)
Source of truth, from cursor/plugins.
name: thermos description: "Launch both thermo-nuclear review subagents in parallel, then synthesize their findings. Use for thermos, double thermo review, or combined bug/security and code-quality branch audits." disable-model-invocation: true
Thermos
Run the two thermo review passes as async background subagents in parallel, then synthesize their results.
Workflow
- Determine the review scope from the user request, PR, current branch, or relevant changed files.
- Gather the diff and any file/context excerpts needed for reviewers to evaluate the change without guessing.
- Launch both subagents in the same message with
run_in_background: true:subagent_type: "thermo-nuclear-review-subagent"for bugs, breakages, security, devex regressions, feature-flag leaks, and other branch-audit risks.subagent_type: "thermo-nuclear-code-quality-review-subagent"for maintainability, structure, file-size growth, spaghetti, abstractions, and codebase-health risks.
- Pass each subagent the same scoped diff/file context and ask it to return prioritized findings with file references and evidence.
- After both finish, synthesize the results with findings first, deduplicated across reviewers. Weight overlapping findings more heavily, resolve disagreements with your own judgment, and keep summaries brief.
If individual background summaries are already visible to the user, do not restate them wholesale. Surface the unified verdict, the highest-signal findings, and any remaining uncertainty.
Related skills
More from cursor/plugins and the wider catalog.
typescript-best-practices
TypeScript best practices for type safety, discriminated unions, and constructive modeling.
unslop
Remove AI writing patterns from any text—detect and fix superficial phrases, vague language, and filler.
verify-this
Verify claims with repeatable local evidence: baseline vs. treatment comparison with VERIFIED, NOT VERIFIED, or INCONCLUSIVE verdict.
weekly-review
Generate a weekly work summary from your commits, categorized by bugfix, tech debt, and new features.
what-did-i-get-done
Summarize your git commits into concise status updates for any time period.
why
Investigate design rationale, tradeoffs, and decision history behind code by querying multiple evidence sources in parallel.