pdf-review-comments

name	pdf-review-comments
description	Extract olive-highlighted text blocks and their review comments from a reviewed PDF into a "Detailed Comments on the Text" markdown section. Use when the user reviewed a paper by highlighting text in olive and attaching comments, and wants those highlights plus comments pulled out of the PDF.
allowed-tools	Bash(jbang *)

PDF Review Comments

Extract olive-highlighted text and its review comments from a PDF and render them as markdown.

Prerequisite

jbang must be installed (jbang --version). It fetches the required JDK and dependencies (Apache PDFBox, aesh) automatically on first run.

How to run

Run the bundled jbang script against the user's PDF:

jbang "${CLAUDE_SKILL_DIR}/PdfReviewComments.java" "<path/to/input.pdf>"

By default this writes <input>.md next to the PDF and prints how many comments were found.

Options (pass after the script, before or around the PDF path):

• -o, --out <file> — write markdown to a specific file instead of the sibling .md.
• -c, --color <RRGGBB> — match a different highlight color (default olive 808000).
• -t, --tolerance <n> — per-channel color match tolerance 0..1 (default 0.15).
• -h, --help — show usage.

Example — capture to stdout for inline display instead of a sibling file:

jbang "${CLAUDE_SKILL_DIR}/PdfReviewComments.java" -o /dev/stdout "paper.pdf"

After running

Report where the markdown was written and, if helpful, show the extracted findings. Each finding is formatted as:

## Detailed Comments on the Text

Page: 1: "the highlighted text"

the reviewer comment

---

Notes

• Only PDF Highlight annotations are read; olive-colored font text (no annotation) is ignored.
• Words split by a line-wrap hyphen are rejoined (compre- hensive → comprehensive); real compounds like well-defined are kept.
• See README.md for full behavior and limitations.