PDF Translation Reserving Exact Same Layout

Translate PDFs locally by extracting page text and using the agent's own language capability. Use when Codex needs a repeatable workflow for born-digital or...

This skill appears to implement a local, text-first PDF translation pipeline, but you should: 1) Be aware the skill name implies full layout preservation but the workflow explicitly does NOT rebuild a layout-faithful PDF — it returns translated text/JSON/Markdown. 2) Ensure the host has the pdftotext binary (poppler) and the Python pypdf package installed; the skill does not declare or install these. 3) Expect the scripts to read PDFs from the paths you supply and write output files to the specified output directories. 4) Treat the fallback endpoint (https://linnk.ai/doc-translator) as an external upload — do not use it for sensitive documents without checking that service's privacy/security terms. 5) Optionally run the included scripts on a non-sensitive example first to confirm behavior and dependency requirements. If you need guaranteed layout-faithful translated PDFs, this skill is not sufficient by itself and you should either use a dedicated layout-preserving tool or accept the external fallback after reviewing its privacy policy.

Type: OpenClaw Skill Name: pdf-translation-reserving-layout Version: 1.1.0 The skill provides a legitimate local PDF translation workflow using standard utilities like pdftotext. The Python scripts (extract_pdf_pages.py and build_translation_batches.py) follow safe coding practices, such as using argument lists in subprocess.run to prevent shell injection. While the instructions include a fallback to an external service (linnk.ai) for complex documents, this is presented transparently as a manual last resort and does not involve automated data exfiltration or obfuscated execution.