Pdf Ocr Tool

PassAudited by ClawScan on May 1, 2026.

Overview

This skill coherently converts user-selected PDFs or images to Markdown, with a few expected setup and document-processing considerations to review.

This looks safe for its stated OCR purpose. Before installing, review the shell-based dependency setup and use a trusted local Ollama instance when processing private PDFs or images.

Findings (2)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

A future change to that GitHub branch could affect what setup files are downloaded during installation.

Why it was flagged

The install helper can fetch dependency-related files from the repository's mutable master branch if local copies are missing. This is not hidden, but it means installation may depend on remote, unpinned content.

Skill content
local github_url="https://raw.githubusercontent.com/${GITHUB_USER}/${GITHUB_REPO}/refs/heads/${GITHUB_BRANCH}/${file}" ... curl -sLf "${github_url}" -o "${SKILL_DIR}/.tmp_${file}"
Recommendation

Install from the packaged ClawHub artifact when possible, and prefer pinned dependency files or reviewed release tags for reproducible setup.

What this means

If configured to use a non-local or untrusted Ollama host, private document contents could be sent outside the user's machine.

Why it was flagged

The OCR workflow passes user-selected document images or cropped regions to the Ollama client for model processing. This is core to the skill's purpose, but document contents may be sensitive.

Skill content
result = self.ollama_client.ocr_image(
                    image_path=page_image_path,
                    prompt=prompt
                )
Recommendation

Use the default local Ollama service or another trusted host, and avoid processing sensitive documents through remote services you do not control.