WiseDiag MedOCR

- Skill renamed from wisediag-medocr to wiseocr. - Expanded and clarified privacy and sensitive data warnings, including a detailed table and lists of prohibited data types. - Updated API key acquisition link for clarity. - Description updated for improved accuracy and readability; no functional changes to usage or features.

- Updated SKILL.md metadata format: moved env_vars block outside credentials for consistency. - No functional or feature changes in this release.

- No functional changes in this version. - Internal structure of SKILL.md updated: the "credentials" block was moved outside the "registry" section. - All guides, features, and usage instructions remain unchanged.

- No changes detected in this version.

- Added support for image OCR: now accepts single image files (jpg, jpeg, png, webp, gif, bmp, tiff) in addition to PDFs. - Updated usage instructions to reflect support for both PDF and image input. - Revised description for broader file compatibility (PDF & Image OCR). - Clarified input argument details to specify that only a single file is supported.

- No changes detected in this version. - All features, documentation, and privacy guidance remain the same.

Version 1.0.20 of wisediag-medocr - No file changes detected; documentation and functionality remain unchanged. - Privacy warnings, usage instructions, and API details are consistent with the previous release.

- Skill renamed from "wiseocr" to "wisediag-medocr" - Shortened and clarified the description for general usage and feature highlights - Updated API key link to a new URL format - Added instruction to never call the API or HTTP endpoints directly; use only the provided script - Removed specific API endpoint details from documentation for better abstraction and safety - General wording and formatting improvements for clarity

- Renamed the skill from wisediag-medocr to wiseocr. - Updated internal references to use the new skill name. - No changes to core functionality or arguments.

- No changes detected in this version; documentation and functionality remain the same.

- Updated the skill description to clarify credential requirements and cloud upload behavior. - Documentation now specifies the API endpoint used for OCR processing. - Instructions emphasize using the provided script exclusively for authentication and file handling. - Improved clarity around data privacy and file handling processes.

Version 1.0.15 of wisediag-medocr - No file changes detected in this release. - Functionality and documentation remain unchanged from the previous version.

- Shortened and clarified the skill description for easier understanding. - Added explicit usage instructions: "Upload a PDF and say Use WiseOCR to OCR this." - Provided a direct API key registration link with improved formatting. - Updated guidance to specifically instruct users never to call APIs or HTTP endpoints directly—scripts must be used. - Removed extra tracking parameters from URLs for cleaner presentation.

No changes detected in this version. - No file changes were found between versions. - Functionality, documentation, and configuration remain the same.

- Updated the API key registration link to include UTM tracking parameters for improved referral tracking. - No other changes detected.

- Updated documentation in SKILL.md and README.md for clarity, including clearer instructions and privacy warnings. - Service endpoint information and documentation link added to SKILL.md. - No changes to core functionality or API usage.

No changes detected in this version. - No file modifications, updates, or new features were introduced. - All documentation and configuration remain the same as in the previous release.

## v1.0.9 — Security & Metadata Hardening ### 🔐 Credential Declaration (IMPORTANT) This skill requires **one** environment variable to function: - `WISEDIAG_API_KEY` — Required. Used to authenticate PDF uploads to the WiseDiag OCR API. No other credentials, tokens, or secrets are needed or accessed. The `registry.credentials` block in SKILL.md explicitly declares: - `required: true` - `env_vars: [WISEDIAG_API_KEY]` If the platform registry metadata does not reflect this requirement, it is a sync/publishing issue — the source of truth is SKILL.md. ### 🛡️ Security Improvements - **Removed `--service-url` CLI parameter:** The API endpoint is now hardcoded to `https://openapi.wisediag.com`. The script can no longer be redirected to arbitrary endpoints, eliminating the risk of unintended data exfiltration. - **Added prominent Privacy Warning:** A ⚠️ Privacy Warning section is now the first thing users see in SKILL.md, clearly stating that files are uploaded to WiseDiag's cloud servers. - **Explicit data flow documentation:** The Data Privacy section now lists exactly what happens to uploaded files in a numbered sequence. ### 📋 What This Skill Does (and Does NOT Do) ✅ Reads WISEDIAG_API_KEY from environment (single credential) ✅ Uploads user-provided PDF to https://openapi.wisediag.com/v1/ocr/pdf ✅ Saves returned Markdown to ~/.openclaw/workspace/WiseOCR/ ✅ Only depends on `requests` (via requirements.txt) ❌ Does NOT read any other environment variables or credentials ❌ Does NOT access files outside the user-provided input PDF ❌ Does NOT allow overriding the API endpoint (hardcoded) ❌ Does NOT request elevated privileges or persistent access ❌ Does NOT modify other skills or system configurations

- Updated API key setup instructions to use a new, shortened link for obtaining your key. - Strengthened warnings to never call the OCR API directly; emphasized usage of the provided script only. - No changes to code or functionality. Documentation update only.

- Updated API key setup instructions to use the direct URL (removed redirect link). - Clarified usage instructions: explicitly state that files must be processed with the provided script, not direct API calls. - No functional or code changes in this release.