Super Ocr
ReviewAudited by ClawScan on May 10, 2026.
Overview
Prompt-injection indicators were detected in the submitted artifacts (unicode-control-chars); human review is required before treating this skill as clean.
This looks like a normal local OCR skill. Before installing, review and preferably pin the Python dependencies, use an isolated environment, and remember that OCR output may expose the contents of the images you choose to process to your agent conversation. ClawScan detected prompt-injection indicators (unicode-control-chars), so this skill requires review even though the model response was benign.
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.
Installing the skill's dependencies may bring in large third-party packages whose exact versions are not fixed by the artifacts.
The helper can install OCR dependencies from package repositories, and the package names are not version-pinned. This is expected for an OCR tool but still creates normal dependency supply-chain exposure.
'install_cmd': 'pip install paddleocr paddlepaddle' ... subprocess.run(install_cmd.split(), capture_output=True, text=True, timeout=120)
Install in a virtual environment, review dependency versions, and pin or audit packages if using this in a sensitive environment.
Using the macOS Vision engine runs local code from the skill package, not just pure Python library calls.
The macOS OCR path executes the included Swift helper with the selected image path. This local subprocess execution is disclosed and purpose-aligned for using Apple's Vision OCR.
subprocess.run(['swift', str(SWIFT_SCRIPT_PATH), image_path], capture_output=True, text=True, timeout=60)
Use the skill only from a trusted copy of the package, and review the local helper script if enabling the MacVision engine.