Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Prevent AI over-engineering by keeping changes scoped, simple, and directly tied to the user's request
.claude/skills/amariahak-prevent-ai-over-engineering-by-keeping-changes-scoped-simple-and-directly-tied/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -36% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-04 | ✓→✗ | ▼ Worse | -63% | 0% |
| case-05 | ✓→✗ | ▼ Worse | -59% | 0% |
| case-16 | ✓→✗ | ▼ Worse | 63% | 0% |
Prevent AI over-engineering by keeping changes scoped, simple, and directly tied to the user's request
Synced from https://github.com/PatrickJS/awesome-cursorrules/tree/main/rules/anti-overengineering.mdc.
Only change what was asked. Simplest solution first. When unsure, ask.
Do not modify unrequested code, add abstractions without a concrete need, import unnecessary dependencies, rewrite entire files for small changes, or add error handling for impossible scenarios.
Before delivery: verify you only changed requested code, check for simpler approaches, confirm no unrequested files were touched.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 6,931 | 4,021 | -42% | 1 | 1 | 0% | 1,394 | 888 | -36% | 0 | 0 | — |
case-02 | fail→pass | 6,310 | 3,568 | -43% | 1 | 1 | 0% | 1,228 | 834 | -32% | 0 | 0 | — |
case-03 | pass→pass | 28,233 | 7,206 | -74% | 1 | 1 | 0% | 6,197 | 1,689 | -73% | 0 | 0 | — |
case-04 | pass→fail | 29,967 | 9,442 | -68% | 1 | 1 | 0% | 5,142 | 1,881 | -63% | 0 | 0 | — |
case-05 | pass→fail | 9,956 | 4,637 | -53% | 1 | 1 | 0% | 2,307 | 957 | -59% | 0 | 0 | — |
case-06 | pass→pass | 3,134 | 2,018 | -36% | 1 | 1 | 0% | 524 | 569 | +9% | 0 | 0 | — |
case-07 | pass→pass | 1,675 | 1,375 | -18% | 1 | 1 | 0% | 244 | 347 | +42% | 0 | 0 | — |
case-08 | pass→pass | 2,962 | 2,225 | -25% | 1 | 1 | 0% | 437 | 450 | +3% | 0 | 0 | — |
case-09 | pass→pass | 3,827 | 1,996 | -48% | 1 | 1 | 0% | 707 | 524 | -26% | 0 | 0 | — |
case-10 | pass→pass | 3,825 | 2,093 | -45% | 1 | 1 | 0% | 319 | 440 | +38% | 0 | 0 | — |
case-11 | pass→pass | 3,386 | 1,971 | -42% | 1 | 1 | 0% | 597 | 470 | -21% | 0 | 0 | — |
case-12 | pass→pass | 3,105 | 2,858 | -8% | 1 | 1 | 0% | 619 | 747 | +21% | 0 | 0 | — |
case-13 | pass→pass | 2,756 | 1,721 | -38% | 1 | 1 | 0% | 481 | 454 | -6% | 0 | 0 | — |
case-14 | pass→pass | 2,139 | 2,201 | +3% | 1 | 1 | 0% | 376 | 529 | +41% | 0 | 0 | — |
case-15 | pass→pass | 2,101 | 2,170 | +3% | 1 | 1 | 0% | 308 | 545 | +77% | 0 | 0 | — |
case-16 | pass→fail | 1,324 | 1,418 | +7% | 1 | 1 | 0% | 224 | 365 | +63% | 0 | 0 | — |
case-17 | pass→pass | 3,121 | 2,296 | -26% | 1 | 1 | 0% | 571 | 581 | +2% | 0 | 0 | — |
case-18 | pass→pass | 1,753 | 1,756 | +0% | 1 | 1 | 0% | 267 | 438 | +64% | 0 | 0 | — |
case-19 | pass→pass | 3,393 | 3,178 | -6% | 1 | 1 | 0% | 497 | 665 | +34% | 0 | 0 | — |
case-20 | pass→pass | 2,515 | 2,016 | -20% | 1 | 1 | 0% | 439 | 523 | +19% | 0 | 0 | — |
case-21 | pass→pass | 1,872 | 2,769 | +48% | 1 | 1 | 0% | 347 | 474 | +37% | 0 | 0 | — |
case-22 | pass→pass | 2,963 | 2,611 | -12% | 1 | 1 | 0% | 526 | 643 | +22% | 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. The headline lift of -20 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
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.