Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Edit text in existing PDFs via natural-language prompts.
.claude/skills/nousresearch-nano-pdf/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -68% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -82% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -64% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -66% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -59% | 0% |
Edit PDFs using natural-language instructions. Point it at a page and describe what to change. For structural PDF work (merge, split, forms, watermarks, creation), see the pdf skill; for text extraction from scans, see ocr-and-documents.
bash# Install with uv (recommended — already available in Hermes) uv pip install nano-pdf # Or with pip pip install nano-pdf
bashnano-pdf edit <file.pdf> <page_number> "<instruction>"
bash# Change a title on page 1 nano-pdf edit deck.pdf 1 "Change the title to 'Q3 Results' and fix the typo in the subtitle" # Update a date on a specific page nano-pdf edit report.pdf 3 "Update the date from January to February 2026" # Fix content nano-pdf edit contract.pdf 2 "Change the client name from 'Acme Corp' to 'Acme Industries'"
read_file to check file size, or open it)nano-pdf --help for config)| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 7,501 | 4,233 | -44% | 1 | 1 | 0% | 1,604 | 615 | -62% | 0 | 0 | — |
case-02 | fail→fail | 8,420 | 5,639 | -33% | 1 | 1 | 0% | 1,908 | 793 | -58% | 0 | 0 | — |
case-03 | fail→fail | 7,511 | 4,073 | -46% | 1 | 1 | 0% | 1,456 | 588 | -60% | 0 | 0 | — |
case-04 | pass→fail | 5,839 | 4,910 | -16% | 1 | 1 | 0% | 1,396 | 617 | -56% | 0 | 0 | — |
case-05 | pass→pass | 8,006 | 7,466 | -7% | 1 | 1 | 0% | 1,740 | 1,977 | +14% | 0 | 0 | — |
case-06 | pass→fail | 10,641 | 7,135 | -33% | 1 | 1 | 0% | 2,460 | 1,034 | -58% | 0 | 0 | — |
case-07 | pass→fail | 7,267 | 4,435 | -39% | 1 | 1 | 0% | 1,590 | 621 | -61% | 0 | 0 | — |
case-08 | fail→pass | 10,283 | 1,387 | -87% | 1 | 1 | 0% | 2,004 | 643 | -68% | 0 | 0 | — |
case-09 | fail→fail | 12,509 | 4,220 | -66% | 1 | 1 | 0% | 2,738 | 652 | -76% | 0 | 0 | — |
case-10 | fail→pass | 15,080 | 1,227 | -92% | 1 | 1 | 0% | 2,754 | 502 | -82% | 0 | 0 | — |
case-11 | fail→pass | 9,385 | 1,617 | -83% | 1 | 1 | 0% | 1,748 | 623 | -64% | 0 | 0 | — |
case-12 | pass→pass | 4,963 | 2,790 | -44% | 1 | 1 | 0% | 999 | 938 | -6% | 0 | 0 | — |
case-13 | pass→pass | 8,159 | 2,445 | -70% | 1 | 1 | 0% | 1,266 | 757 | -40% | 0 | 0 | — |
case-14 | pass→pass | 8,280 | 2,990 | -64% | 1 | 1 | 0% | 1,816 | 950 | -48% | 0 | 0 | — |
case-15 | pass→pass | 16,054 | 5,752 | -64% | 1 | 1 | 0% | 2,262 | 1,421 | -37% | 0 | 0 | — |
case-16 | fail→pass | 8,233 | 1,446 | -82% | 1 | 1 | 0% | 1,692 | 573 | -66% | 0 | 0 | — |
case-17 | fail→pass | 9,993 | 2,812 | -72% | 1 | 1 | 0% | 2,154 | 882 | -59% | 0 | 0 | — |
case-18 | fail→pass | 11,033 | 1,541 | -86% | 1 | 1 | 0% | 2,349 | 655 | -72% | 0 | 0 | — |
case-19 | fail→pass | 8,504 | 2,881 | -66% | 1 | 1 | 0% | 1,813 | 571 | -69% | 0 | 0 | — |
case-20 | fail→pass | 9,177 | 1,262 | -86% | 1 | 1 | 0% | 1,928 | 547 | -72% | 0 | 0 | — |
case-21 | pass→pass | 3,350 | 2,066 | -38% | 1 | 1 | 0% | 702 | 734 | +5% | 0 | 0 | — |
case-22 | fail→pass | 7,369 | 1,458 | -80% | 1 | 1 | 0% | 1,607 | 606 | -62% | 0 | 0 | — |
case-23 | fail→fail | 8,760 | 3,787 | -57% | 1 | 1 | 0% | 1,891 | 604 | -68% | 0 | 0 | — |
case-24 | pass→pass | 3,647 | 1,555 | -57% | 1 | 1 | 0% | 777 | 645 | -17% | 0 | 0 | — |
case-25 | fail→pass | 4,127 | 1,564 | -62% | 1 | 1 | 0% | 896 | 604 | -33% | 0 | 0 | — |
case-26 | fail→pass | 5,438 | 2,299 | -58% | 1 | 1 | 0% | 1,122 | 528 | -53% | 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. 26 cases were attempted, and 18 counted toward the lift figure. The other 8 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 +31 percentage points is the difference between those two pass rates over the 18 comparable cases. 5 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.