Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Installs and runs Million's React Doctor to audit a React or Next.js codebase for performance, security, correctness, and architecture issues. Use when the user mentions react-doctor, React Doctor, millionco/react-doctor, performance audits, code health, bundle size, or suspicious React patterns.
.claude/skills/asymmetric-al-react-doctor/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -36% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -46% | 0% |
Run an automated audit of a React/Next.js codebase and turn the diagnostics into a prioritized fix list.
In this repository, use the repo-owned first-party helper from the repo root:
bashbun run react-doctor:first-party -- --full --offline --fail-on none
The human guide is docs/guides/development/react-doctor.md. It is the source of truth for configured ignores, advisory blocking behavior, and PR wording.
package.json).docs/guides/development/react-doctor.md.bun run react-doctor:first-party -- --full --offline --fail-on noneapps and packages.When triaging results for Next.js App Router projects, prioritize:
"use client" leaks)metadata exportsWhen reporting results back, format as:
markdownReact Doctor score: <number>/100 Top findings: - <severity> <short title> -- <path(s)> - ... Fix plan: 1. <most important fix> (why) 2. ... Verification: - bun run react-doctor:first-party -- --full --offline --fail-on none
docs/guides/development/react-doctor.mdbun run react-doctor:first-party -- --full --offline --fail-on noneblocking behavior reported honestly| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 17,813 | 5,284 | -70% | 1 | 1 | 0% | 3,150 | 963 | -69% | 0 | 0 | — |
case-02 | fail→fail | 16,345 | 6,040 | -63% | 1 | 1 | 0% | 2,915 | 962 | -67% | 0 | 0 | — |
case-03 | fail→fail | 11,600 | 5,717 | -51% | 1 | 1 | 0% | 1,934 | 936 | -52% | 0 | 0 | — |
case-04 | fail→pass | 10,957 | 2,754 | -75% | 1 | 1 | 0% | 1,734 | 1,110 | -36% | 0 | 0 | — |
case-05 | fail→pass | 12,299 | 2,171 | -82% | 1 | 1 | 0% | 2,012 | 987 | -51% | 0 | 0 | — |
case-06 | pass→pass | 7,910 | 3,987 | -50% | 1 | 1 | 0% | 1,213 | 1,212 | -0% | 0 | 0 | — |
case-07 | pass→pass | 10,416 | 9,369 | -10% | 1 | 1 | 0% | 1,793 | 2,211 | +23% | 0 | 0 | — |
case-08 | fail→pass | 13,684 | 7,037 | -49% | 1 | 1 | 0% | 2,108 | 1,751 | -17% | 0 | 0 | — |
case-09 | pass→pass | 8,278 | 4,860 | -41% | 1 | 1 | 0% | 1,296 | 1,430 | +10% | 0 | 0 | — |
case-10 | pass→pass | 10,083 | 2,533 | -75% | 1 | 1 | 0% | 1,557 | 1,091 | -30% | 0 | 0 | — |
case-11 | pass→pass | 10,643 | 8,635 | -19% | 1 | 1 | 0% | 1,890 | 2,116 | +12% | 0 | 0 | — |
case-12 | pass→pass | 6,752 | 2,028 | -70% | 1 | 1 | 0% | 993 | 959 | -3% | 0 | 0 | — |
case-13 | pass→pass | 8,175 | 3,843 | -53% | 1 | 1 | 0% | 1,378 | 1,248 | -9% | 0 | 0 | — |
case-14 | fail→pass | 9,229 | 4,574 | -50% | 1 | 1 | 0% | 1,406 | 1,442 | +3% | 0 | 0 | — |
case-15 | pass→pass | 14,656 | 9,101 | -38% | 1 | 1 | 0% | 2,352 | 2,171 | -8% | 0 | 0 | — |
case-16 | fail→pass | 10,220 | 1,844 | -82% | 1 | 1 | 0% | 1,787 | 964 | -46% | 0 | 0 | — |
case-17 | fail→pass | 9,305 | 5,618 | -40% | 1 | 1 | 0% | 1,482 | 1,582 | +7% | 0 | 0 | — |
case-18 | pass→pass | 6,891 | 5,322 | -23% | 1 | 1 | 0% | 1,278 | 1,570 | +23% | 0 | 0 | — |
case-19 | pass→pass | 11,273 | 4,223 | -63% | 1 | 1 | 0% | 1,739 | 1,390 | -20% | 0 | 0 | — |
case-20 | pass→pass | 10,687 | 7,511 | -30% | 1 | 1 | 0% | 2,012 | 2,028 | +1% | 0 | 0 | — |
case-21 | pass→pass | 8,872 | 6,088 | -31% | 1 | 1 | 0% | 1,938 | 1,911 | -1% | 0 | 0 | — |
case-22 | pass→pass | 7,252 | 3,251 | -55% | 1 | 1 | 0% | 1,478 | 1,197 | -19% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +27 percentage points is the difference between those two pass rates over the 19 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.