Cofferdock

Each tool maps to a distinct operation: PDF transformations, format conversions, validations, and generators are cleanly separated, and the few similar pairs (html_to_image vs screenshot_url, protect_pdf vs unprotect_pdf) are differentiated by input and direction. There is no real risk of an agent selecting the wrong tool for a task.

All names use snake_case and most fall into predictable groups: verb_noun (merge_pdfs, rotate_pdf), source_to_target (pdf_to_images, html_to_pdf), and validate_* (validate_email, validate_iban, validate_phone). The minor inconsistency comes from noun-only or verb-only names like pdf_metadata, hash, and slugify, so the overall pattern is readable but not perfectly uniform.

At 26 tools, the server sits just above the heavy range, and the broad utility scope means no tool is actually redundant. Still, the number is larger than an agent typically needs, and several one-off utilities like slugify and generate_uuid could arguably be bundled or omitted.

The PDF workflow is remarkably complete: create, extract, fill, merge, split, rotate, watermark, protect/unprotect, metadata, and page numbering are all covered, with image, HTML, and validation utilities rounding it out. Minor gaps exist—no unzip counterpart to create_zip, no PDF compression, and no OCR—but they are either documented or outside the core workflows.