{"data":[{"skill_name":"summarize","skill_description":"Summarize or transcribe URLs, YouTube/videos, podcasts, articles, transcripts, PDFs, and local files.","author":"openclaw","stars":372728,"skill_url":"https://github.com/openclaw/openclaw/blob/1300b2263027f1f9a72a2e5fbde2fd272536461f/skills/summarize","category":"AIGC","evaluation":{"safety":{"level":"Good","reason":"The skill is a benign content summarization tool that processes user-provided URLs, files, and videos. It does not perform destructive actions, bypass security, or encourage unsafe tool usage. The use of external LLM APIs is standard for this type of tool and does not constitute a safety risk."},"completeness":{"level":"Good","reason":"The SKILL.md provides comprehensive documentation including trigger phrases, quick start examples, YouTube summary vs transcript modes, required API keys for multiple providers, useful flags (--length, --extract, --json, etc.), and configuration file examples. All essential steps, inputs, and prerequisites are clearly specified."},"executability":{"level":"Good","reason":"The skill provides concrete, executable bash commands (e.g., 'summarize https://example.com', 'summarize /path/to/file.pdf') with specific flags. The workflow is clear and actionable - an agent can directly execute these commands. No ambiguous or hand-wavy steps are present."},"cost_awareness":{"level":"Good","reason":"The task is inherently lightweight (single content summarization calls). The skill explicitly includes cost-control mechanisms: --length flag (short/medium/long/xl/xxl), --max-output-tokens parameter, and the ability to choose different providers/models. No heavy loops, large-scale processing, or unbounded operations are suggested."},"maintainability":{"level":"Good","reason":"The skill has a narrow, modular scope focused solely on summarization. It offers multiple configuration points (config file, environment variables for different API providers) and adjustable parameters (--length, --model). The skill is easily reusable and can be composed with other workflows without tight coupling."}}},{"skill_name":"nano-pdf","skill_description":"Edit PDFs with natural-language instructions using the nano-pdf CLI.","author":"clawdbot","stars":230194,"skill_url":"https://github.com/clawdbot/clawdbot/blob/02aeff8efcd604bbab0effba3ed71b247ff00ebb/skills/nano-pdf","category":"Productivity","evaluation":{"safety":{"level":"Good","reason":"The Skill describes a benign CLI operation (natural-language edits to a PDF) and explicitly advises users to \"sanity-check the output PDF before sending it out,\" with no instructions encouraging destructive or unsafe actions."},"completeness":{"level":"Average","reason":"The SKILL.md gives a clear example command (\"nano-pdf edit deck.pdf 1 ...\") and a note about page-number ambiguity, but it omits details like whether edits are in-place or written to a new file, error handling, supported PDF features, and other CLI options."},"executability":{"level":"Good","reason":"The workflow is concrete (a runnable CLI example is provided) and the metadata lists the required binary (\"bins\": [\"nano-pdf\"]) so an agent with the tool installed can execute it, though page-number indexing may require retry as noted."},"cost_awareness":{"level":"Good","reason":"Editing a PDF via a local CLI is inherently low-cost and the documentation contains only a short example and notes, with no indications of heavy compute or external service usage."},"maintainability":{"level":"Average","reason":"The Skill has a narrow, modular scope (single CLI command) making it easy to adapt by changing arguments, but SKILL.md lacks configuration points and detailed options, which limits straightforward extension or integration."},"script_execution":[]}},{"skill_name":"summarize","skill_description":"Summarize URLs or files with the summarize CLI (web, PDFs, images, audio, YouTube).","author":"clawdbot","stars":230194,"skill_url":"https://github.com/clawdbot/clawdbot/blob/920b3880c1465e5978041dcdd2a39018b9f620db/skills/summarize","category":"Research","evaluation":{"safety":{"level":"Good","reason":"The skill is a benign summarization CLI and contains no instructions for destructive actions or credential exfiltration; metadata only requires the 'summarize' binary and examples focus on read-only summarization of URLs/files (SKILL.md and metadata)."},"completeness":{"level":"Good","reason":"The SKILL.md gives concrete usage examples, required API key environment variables, install instructions (brew formula in metadata), and useful flags (length, --extract-only, --json, fallbacks), covering the essential inputs and options needed to run the tool."},"executability":{"level":"Good","reason":"The README/SKILL.md provides explicit CLI examples and an install path (brew steipete/tap/summarize) plus required environment variables for providers, so an agent with typical shell access could install and run the commands as documented."},"cost_awareness":{"level":"Good","reason":"The documentation includes explicit controls for output size and model usage (--length, --max-output-tokens, configurable model via config file) which allow users to limit token usage when calling large language models."},"maintainability":{"level":"Good","reason":"The tool exposes many configurable flags, a ~/.summarize/config.json for default model selection, and supports multiple providers and fallbacks (FIRECRAWL_API_KEY, APIFY_API_TOKEN), making it easy to adapt behavior and integrate into other workflows."},"script_execution":[]}},{"skill_name":"visa-doc-translate","skill_description":"将签证申请文件（图片）翻译成英文，并创建包含原文和译文的双语PDF","author":"affaan-m","stars":128071,"skill_url":"https://github.com/affaan-m/everything-claude-code/blob/bfc73866c92b4c2dde4b3639c736b80483b4c195/docs/zh-CN/skills/visa-doc-translate","category":"Productivity","evaluation":{"safety":{"level":"Good","reason":"The skill describes benign operations (image conversion, OCR, translation, PDF generation) and does not include destructive actions or instructions to bypass security; the SKILL.md explicitly restricts behavior to processing documents and producing a PDF."},"completeness":{"level":"Average","reason":"The SKILL.md gives a full high-level workflow (HEIC conversion, EXIF-based rotation, OCR fallback order, translation rules, PDF layout and output naming) but omits precise implementation details for key steps (exact rotation-detection/‘test 180 degrees’ algorithm, OCR orchestration logic, and no actual Python scripts shown)."},"executability":{"level":"Average","reason":"The README and SKILL.md provide concrete commands and library requirements (sips, pip installs, PIL/reportlab, OCR options) making the workflow actionable, but there are no provided runnable scripts and several steps (rotation heuristics, Vision framework usage examples, PDF generation code) are underspecified."},"cost_awareness":{"level":"Good","reason":"The described approach uses local OCR libraries (EasyOCR/Tesseract or macOS Vision) and standard Python libraries (Pillow, reportlab) without heavy external APIs or large-scale processing, indicating low compute and monetary cost."},"maintainability":{"level":"Average","reason":"The skill has a narrow, well-scoped purpose and clear input/output naming, but it is somewhat coupled to platform-specific options (macOS Vision vs EasyOCR/Tesseract) and lacks configuration points or modular example code, making adaptation require non-trivial implementation work."}}},{"skill_name":"nutrient-document-processing","skill_description":"Process, convert, OCR, extract, redact, sign, and fill documents using the Nutrient DWS API. Works with PDFs, DOCX, XLSX, PPTX, HTML, and images.","author":"affaan-m","stars":128071,"skill_url":"https://github.com/affaan-m/everything-claude-code/blob/3bc8672432d628178ea7b5a617b642ea4bf07e38/skills/nutrient-document-processing","category":"Productivity","evaluation":{"safety":{"level":"Average","reason":"The skill involves processing sensitive documents and redacting PII, but lacks explicit safeguards like user confirmation for destructive actions or warnings about data privacy risks, though it does not encourage credential exfiltration or system damage."},"completeness":{"level":"Good","reason":"SKILL.md clearly describes multiple operations with concrete examples, prerequisites like API key setup, and links to full documentation, covering essential steps and constraints for document processing."},"executability":{"level":"Good","reason":"The skill provides specific curl commands and an MCP server alternative with clear instructions, making it actionable with typical tools, and it is designed as an instruction-only skill that does not require runnable scripts."},"cost_awareness":{"level":"Average","reason":"The skill uses an external API which may incur costs, but it does not explicitly mention batching, limits, or cost control measures, though the examples suggest lightweight usage without obviously wasteful behavior."},"maintainability":{"level":"Good","reason":"The skill has a narrow, modular scope focused on document processing via an API, with configurable parameters in instructions and clear inputs/outputs, allowing easy adaptation and combination with other skills."}}},{"skill_name":"nutrient-document-processing","skill_description":"Nutrient DWS API を使用してドキュメントの処理、変換、OCR、抽出、編集、署名、フォーム入力を行います。PDF、DOCX、XLSX、PPTX、HTML、画像に対応しています。","author":"affaan-m","stars":128071,"skill_url":"https://github.com/affaan-m/everything-claude-code/blob/277006bd7f18924ce9687b7546f7e83c9c74ba9c/docs/ja-JP/skills/nutrient-document-processing","category":"Productivity","evaluation":{"safety":{"level":"Average","reason":"The skill involves document processing with PII redaction capabilities but lacks explicit safety disclaimers about handling sensitive data or confirmation prompts for destructive actions like redaction, though the operations appear benign overall."},"completeness":{"level":"Good","reason":"SKILL.md provides comprehensive documentation with clear goals, detailed API usage examples for all advertised capabilities, prerequisites (API key setup), and links to external documentation for additional details."},"executability":{"level":"Good","reason":"The skill provides concrete curl commands with specific parameters and file handling, and offers an alternative MCP server setup; the workflow is actionable with typical tools despite being instruction-only."},"cost_awareness":{"level":"Average","reason":"The skill uses an external API service which inherently involves costs, but SKILL.md does not mention rate limits, pricing tiers, or optimization strategies like batching to control operational expenses."},"maintainability":{"level":"Good","reason":"The skill has a narrow, modular scope focused on document processing via a well-defined API, with clear configuration points (API key, instructions JSON) and low coupling to specific implementations."}}},{"skill_name":"nutrient-document-processing","skill_description":"使用Nutrient DWS API处理、转换、OCR识别、提取、编辑、签名和填写文档。支持PDF、DOCX、XLSX、PPTX、HTML和图像格式。","author":"affaan-m","stars":128071,"skill_url":"https://github.com/affaan-m/everything-claude-code/blob/af51fcacb70d42ed52ba8111026d9b0f0e49b364/docs/zh-CN/skills/nutrient-document-processing","category":"Productivity","evaluation":{"safety":{"level":"Good","reason":"The skill operates on document processing tasks (conversion, OCR, redaction, watermark, signing, form fill) and provides non-destructive curl examples and an API-key setup, without instructions to perform harmful actions or credential exfiltration (e.g., SKILL.md shows redaction, watermark, sign examples and env var NUTRIENT_API_KEY)."},"completeness":{"level":"Good","reason":"The README includes setup (exporting NUTRIENT_API_KEY), the endpoint (https://api.nutrient.io/build), concrete multipart curl examples for conversion, OCR, extraction, redaction, watermark, signing and form filling, lists supported formats and OCR languages, and links to full docs and playground."},"executability":{"level":"Good","reason":"Multiple concrete curl commands with required headers, file fields, and instructions JSON are provided (e.g., DOCX->PDF, OCR to searchable PDF, redaction, fillForm), so an agent with an API key and files can execute the described workflows."},"cost_awareness":{"level":"Average","reason":"The SKILL.md mentions obtaining a free API key but does not document rate limits, file size limits, batching, or cost controls for potentially heavy operations like OCR or large-scale extraction."},"maintainability":{"level":"Good","reason":"The Skill uses a clear, modular 'instructions' JSON pattern and shows an MCP server configuration with env vars (NUTRIENT_DWS_API_KEY, SANDBOX_PATH), making it straightforward to adapt or extend for other document actions."}}},{"skill_name":"visa-doc-translate","skill_description":"Translate visa application documents (images) to English and create a bilingual PDF with original and translation","author":"affaan-m","stars":128071,"skill_url":"https://github.com/affaan-m/everything-claude-code/blob/2d3be88bb550c48f1c44ab260934021615dd3cad/skills/visa-doc-translate","category":"Productivity","evaluation":{"safety":{"level":"Good","reason":"The skill only describes processing user-supplied document images (convert/rotate/OCR/translate/generate PDF) and contains no destructive or privileged operations; SKILL.md does not instruct any risky system-modifying actions."},"completeness":{"level":"Average","reason":"The SKILL.md enumerates concrete steps (HEIC conversion, EXIF rotation, OCR options, required libraries, PDF layout) but does not include the actual Python PDF-generation script, a concrete translation implementation or error-handling/failure-mode policies for OCR/translation fallbacks."},"executability":{"level":"Average","reason":"The document lists specific tools and commands (sips conversion, macOS Vision, EasyOCR, Tesseract, pillow/reportlab installs) making the workflow feasible, but no runnable scripts are provided and there are platform-specific dependencies and a potentially incorrect/underspecified EXIF rotation instruction that may cause implementation issues."},"cost_awareness":{"level":"Average","reason":"The described tasks are generally low-cost and local (OCR, image conversion, PDF generation), but the SKILL.md gives no guidance on rate-limiting, batching, or costs for translation APIs or large-scale processing, so operational cost controls are missing."},"maintainability":{"level":"Average","reason":"The skill has a narrow, modular scope and lists dependencies, but contains hard-coded rules (e.g., specific EXIF orientation handling and Chinese-to-pinyin naming convention) and lacks configuration points or sample code structure for easy reuse or extension."}}},{"skill_name":"nutrient-document-processing","skill_description":"Process, convert, OCR, extract, redact, sign, and fill documents using the Nutrient DWS API. Works with PDFs, DOCX, XLSX, PPTX, HTML, and images.","author":"affaan-m","stars":128071,"skill_url":"https://github.com/affaan-m/everything-claude-code/blob/3bc8672432d628178ea7b5a617b642ea4bf07e38/.cursor/skills/nutrient-document-processing","category":"Productivity","evaluation":{"safety":{"level":"Average","reason":"The skill involves processing sensitive documents (e.g., redacting PII, digital signatures) but lacks explicit safeguards like user confirmations or scope limits for destructive actions, though it does not encourage credential exfiltration or system damage."},"completeness":{"level":"Good","reason":"SKILL.md clearly describes setup, operations, and examples for document processing tasks, including prerequisites like API key and supported formats, with no critical correctness errors in the provided instructions."},"executability":{"level":"Good","reason":"The skill provides concrete curl commands and an MCP server alternative with specific parameters and endpoints, making it actionable with typical tools; it is an instruction-only skill that does not require runnable scripts."},"cost_awareness":{"level":"Average","reason":"The skill uses an external API which could incur costs, but it does not explicitly mention batching, limits, or caching to control expenses, though nothing suggests obviously wasteful behavior."},"maintainability":{"level":"Good","reason":"The skill has a modular scope with clearly defined inputs and outputs for each operation (e.g., convert, extract, OCR), and it mentions configuration points like API keys and instruction parameters rather than hard-coding assumptions."}}},{"skill_name":"docx","skill_description":"Use this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files). Triggers include: any mention of \"Word doc\", \"word document\", \".docx\", or requests to produce professional documents with formatting like tables of contents, headings, page numbers, or letterheads. Also use when extracting or reorganizing content from .docx files, inserting or replacing images in documents, performing find-and-replace in Word files, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a \"report\", \"memo\", \"letter\", \"template\", or similar deliverable as a Word or .docx file, use this skill. Do NOT use for PDFs, spreadsheets, Google Docs, or general coding tasks unrelated to document generation.","author":"anthropics","stars":91061,"skill_url":"https://github.com/anthropics/skills/blob/a5bcdd7e58cdff48566bf876f0a72a2008dcefbc/skills/docx","category":"Productivity","evaluation":{"safety":{"level":"Good","reason":"The Skill describes non-destructive document operations (create/read/edit/unpack/repack) using docx-js, pandoc, and local scripts (e.g., scripts/office/unpack.py and scripts/accept_changes.py) and does not instruct credential exfiltration, system-destruction, or other unsafe actions."},"completeness":{"level":"Average","reason":"The SKILL.md is detailed (explicit docx-js rules for page size, tables, images, TOC and an editing workflow) but appears truncated: it says 'Follow all 3 steps' for editing yet only shows Step 1 and Step 2 and some script/templates are incomplete, leaving key repack/finish steps and some implementation details uncertain."},"executability":{"level":"Average","reason":"There are concrete, runnable commands and code examples (docx-js sample, pandoc commands, scripts/office/unpack.py) and many scripts compiled successfully in the execution log, but several files/snippets are truncated and the workflow requires external tooling (e.g., LibreOffice) and missing inputs/Step 3 to be fully executable without extra setup."},"cost_awareness":{"level":"Good","reason":"The described approach uses local, lightweight tools (docx-js, pandoc, LibreOffice) and operates on individual documents, so it does not imply heavy compute or ongoing external costs."},"maintainability":{"level":"Good","reason":"The Skill is modular and configurable with clear parameters and conventions (page size DXA constants, numbering/table rules, styles override, and separate scripts under scripts/), making it straightforward to adapt or extend."},"script_execution":[{"note":"no usage examples found; py_compile succeeded","path":"scripts/__init__.py","error":null,"status":"compiled_only","command":"python -m py_compile scripts/__init__.py","exit_code":0,"duration_sec":0.045},{"note":"no usage examples found; py_compile succeeded","path":"scripts/accept_changes.py","error":null,"status":"compiled_only","command":"python -m py_compile scripts/accept_changes.py","exit_code":0,"duration_sec":0.037},{"note":"missing inputs: unpacked/","path":"scripts/comment.py","error":null,"status":"compiled_only","command":"python -m py_compile scripts/comment.py","exit_code":0,"duration_sec":0.037},{"note":"no usage examples found; py_compile succeeded","path":"scripts/office/helpers/__init__.py","error":null,"status":"compiled_only","command":"python -m py_compile scripts/office/helpers/__init__.py","exit_code":0,"duration_sec":0.034},{"note":"no usage examples found; py_compile succeeded","path":"scripts/office/helpers/merge_runs.py","error":null,"status":"compiled_only","command":"python -m py_compile scripts/office/helpers/merge_runs.py","exit_code":0,"duration_sec":0.034}]}}],"meta":{"query":"pdf","mode":"keyword","category":null,"limit":10,"total":44999,"page":1,"min_stars":0,"sort_by":"stars","sort_order":"desc","threshold":null},"success":true}