Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when a user asks to "proofread", "review and correct", "fix grammar", "improve readability while keeping my voice", and to proofread a document file and save an updated version.
.claude/skills/sickn33-professional-proofreader/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 176% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 156% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 121% | 0% |
| case-06 | ✓→✓ | ＝ Same ✓ | 189% | 0% |
This skill transforms flawed writing — whether pasted text or uploaded documents — into publication-ready prose without altering the author’s intent. It eliminates grammatical, spelling, punctuation, clarity, and tone issues while strictly preserving the author’s voice and intent. Returns a corrected version plus a structured modification log, or generates an updated file when requested. Not for code editing or technical refactoring.
This skill operates in two modes:
Refer markdown for complete inline text mode.
Trigger when user says:
Refer markdown for complete file processing mode.
If inline: -> Return Corrected Version + Modifications list.
If file rewrite: -> Save updated file. -> Confirm filename. -> Provide modifications list unless suppressed.
Give friendly message to user in the end.
User request:
> Proofread this text for grammar, clarity, consistency, and readability while preserving my voice.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 1,994 | 2,394 | +20% | 1 | 1 | 0% | 364 | 1,005 | +176% | 0 | 0 | — |
case-02 | fail→pass | 2,483 | 2,988 | +20% | 1 | 1 | 0% | 439 | 1,126 | +156% | 0 | 0 | — |
case-03 | fail→fail | 8,320 | 6,918 | -17% | 1 | 1 | 0% | 1,632 | 1,779 | +9% | 0 | 0 | — |
case-04 | pass→pass | 4,339 | 4,123 | -5% | 1 | 1 | 0% | 604 | 1,333 | +121% | 0 | 0 | — |
case-05 | fail→pass | 12,929 | 8,528 | -34% | 1 | 1 | 0% | 2,518 | 2,102 | -17% | 0 | 0 | — |
case-06 | pass→pass | 2,496 | 3,151 | +26% | 1 | 1 | 0% | 425 | 1,228 | +189% | 0 | 0 | — |
case-07 | pass→pass | 2,979 | 3,843 | +29% | 1 | 1 | 0% | 481 | 1,237 | +157% | 0 | 0 | — |
case-08 | pass→pass | 5,815 | 3,579 | -38% | 1 | 1 | 0% | 1,000 | 1,269 | +27% | 0 | 0 | — |
case-09 | pass→pass | 3,077 | 2,221 | -28% | 1 | 1 | 0% | 537 | 984 | +83% | 0 | 0 | — |
case-10 | pass→pass | 4,424 | 4,755 | +7% | 1 | 1 | 0% | 824 | 1,428 | +73% | 0 | 0 | — |
case-11 | pass→pass | 3,459 | 3,409 | -1% | 1 | 1 | 0% | 656 | 1,253 | +91% | 0 | 0 | — |
case-12 | pass→pass | 4,254 | 3,492 | -18% | 1 | 1 | 0% | 787 | 1,228 | +56% | 0 | 0 | — |
case-13 | pass→pass | 6,465 | 3,636 | -44% | 1 | 1 | 0% | 1,034 | 1,236 | +20% | 0 | 0 | — |
case-14 | pass→pass | 5,355 | 2,315 | -57% | 1 | 1 | 0% | 1,045 | 1,031 | -1% | 0 | 0 | — |
case-15 | fail→fail | 2,808 | 5,795 | +106% | 1 | 1 | 0% | 404 | 1,562 | +287% | 0 | 0 | — |
case-16 | pass→pass | 4,355 | 3,253 | -25% | 1 | 1 | 0% | 747 | 1,189 | +59% | 0 | 0 | — |
case-17 | pass→pass | 2,124 | 2,025 | -5% | 1 | 1 | 0% | 351 | 915 | +161% | 0 | 0 | — |
case-18 | pass→pass | 3,982 | 4,229 | +6% | 1 | 1 | 0% | 847 | 1,443 | +70% | 0 | 0 | — |
case-19 | pass→pass | 3,168 | 3,212 | +1% | 1 | 1 | 0% | 601 | 1,169 | +95% | 0 | 0 | — |
case-20 | pass→pass | 4,284 | 2,341 | -45% | 1 | 1 | 0% | 838 | 1,046 | +25% | 0 | 0 | — |
case-21 | pass→pass | 3,817 | 2,613 | -32% | 1 | 1 | 0% | 639 | 1,073 | +68% | 0 | 0 | — |
case-22 | fail→fail | 2,704 | 6,295 | +133% | 1 | 1 | 0% | 435 | 1,685 | +287% | 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 +14 percentage points is the difference between those two pass rates over the 22 comparable cases.
The publisher has shipped newer versions since this run, so these numbers describe v1, not the version currently listed.
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.