- Added new dependencies for dask, including compatibility requirements for Python 3.10+ and dask 2025.1+.
- Updated pymoo documentation to reflect version 1.1, including installation instructions and new features for mixed-variable optimization.
- Revised scanpy documentation to version 1.1, incorporating updates for doublet detection, pseudobulk aggregation, and improved autosave functionality.
- Enhanced clarity in examples and troubleshooting sections across all skills to improve user guidance.
- Incremented version numbers to 1.1 for both skills to reflect updates and improvements.
- Revised the scientific-critical-thinking skill to clarify visual aids usage and compatibility requirements, including API access for optional figures.
- Enhanced references in scientific-critical-thinking to include updated risk of bias assessment tools and reporting guidelines.
- Updated stable-baselines3 skill documentation to specify installation instructions, compatibility requirements, and additional resources for users.
- Improved clarity in examples and troubleshooting sections for both skills to enhance user guidance.
- Incremented version number to 1.1 to reflect updates and improvements.
- Added compatibility requirements for Python 3.11+ and scikit-learn 1.7+.
- Updated installation instructions to specify the correct package name and optional dependencies.
- Revised model evaluation examples to use `root_mean_squared_error` instead of `mean_squared_error` for clarity.
- Enhanced preprocessing and supervised learning references to reflect changes in parameter names and deprecations.
- Included new features in dimensionality reduction and clustering sections, such as ClassicalMDS and updates to HDBSCAN.
- Improved clarity and consistency across various documentation sections for better user guidance.
- Enhanced the research grants SKILL.md to include updated compliance with PAPPG 24-1 and NIH salary cap guidance, reflecting changes effective May 20, 2024.
- Revised budget justification template to clarify salary cap details and linked to the NIH Salary Cap Summary for better user reference.
- Updated NIH guidelines to specify the Executive Level II salary cap and its implications for grant proposals.
- Removed outdated schematic generation scripts to streamline the skill's functionality.
- Incremented version number to 1.1 to indicate these updates and improvements.
- Enhanced the histolab SKILL.md with new features including built-in stain normalization and installation requirements for Python 3.8–3.11.
- Updated the astropy SKILL.md to reflect version 1.1, including installation instructions for specific versions and additional notes on network access and reproducibility.
- Improved clarity in the references for both skills, emphasizing network considerations and best practices for data handling.
- Incremented version numbers for both skills to 1.1 to indicate updates and improvements.
- Added a new section encouraging users to star the repository for visibility and support.
- Removed the previous star encouragement to streamline the message and avoid redundancy.
- Added a new section for Star History to showcase project popularity.
- Updated skills usage descriptions to standardize formatting and improve clarity.
- Removed the previous Star History section to streamline the README.
- Added compatibility requirements for the BGPT MCP server and optional API key for paid usage.
- Incremented version number to 1.1 for clarity on updates.
- Revised setup and usage instructions to enhance user guidance on configuring the BGPT MCP server and utilizing the `search_papers` tool.
- Updated the license to MIT and added compatibility requirements for Python 3.10+ and API key usage.
- Incremented version number to 1.2 and included additional metadata for clarity.
- Revised installation instructions and environment variable handling for better user guidance.
- Enhanced documentation with official API links and improved code examples for consistency.
- Enhanced the description of the Benchling integration to clarify the use of the Python SDK and REST API for various registry entities and workflows.
- Updated the license to MIT and added allowed tools for better clarity on usage.
- Incremented the version number to 1.2 and improved installation instructions, including environment variable handling and authentication methods.
- Revised API endpoint references and authentication documentation to reflect best practices and ensure consistency across the integration guide.
- Revised README to reflect the rebranding from Claude Scientific Skills to Scientific Agent Skills, highlighting broader compatibility and the introduction of K-Dense BYOK.
- Updated the number of available skills from 139 to 140 in the README and various SKILL.md files.
- Incremented version numbers for multiple skills to 1.1, ensuring consistency across documentation.
- Enhanced descriptions and allowed tools formatting in several SKILL.md files for improved clarity and usability.
- Added Google Antigravity to the list of compatible agents in multiple sections of the README.
- Included a new section encouraging users to follow K-Dense on social media for updates, tutorials, and new skills announcements.
- Improved clarity and consistency in the documentation regarding supported platforms and installation instructions.
Remove AGENTS.md from version control and ignore it so local agent instructions are not published on GitHub.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Added detailed instructions for adding or updating skills, referencing the CONTRIBUTING.md file for clarity.
- Updated citation section to emphasize the importance of citing both the overall collection and individual skills, including examples for various citation formats.
- Improved clarity on versioning requirements for skills and the importance of acknowledging contributions in publications.
- Added version "1.0" and skill author information to multiple SKILL.md files across various scientific skills.
- Enhanced documentation consistency and clarity by standardizing metadata formatting.
- Ensured all skills now include versioning for better tracking of updates and changes.
- Revised the README and scientific skills documentation to enhance clarity and detail, including updates to the What-If Oracle and Vaex skills.
- Improved descriptions for scenario analysis and data processing capabilities, emphasizing new features and usage examples.
- Updated installation instructions and compatibility notes for Vaex and Zarr skills, ensuring alignment with the latest library versions.
- Enhanced the xlsx skill documentation to clarify requirements and installation steps, including optional dependencies for improved functionality.
- Updated the PyTorch Lightning skill description to reflect the latest version (lightning 2.6+) and clarify installation requirements.
- Changed installation commands from `pip` to `uv pip` for consistency across documentation.
- Removed references to `NeptuneLogger` due to its removal in version 2.6.4, advising users to use alternative loggers instead.
- Enhanced various sections with additional details on compatibility and usage examples for improved clarity.
- Bumped version of 'scientific-agent-skills' from 2.40.0 to 2.42.0 to incorporate recent enhancements and updates.
- Improved documentation for Arboreto and AnnData skills, including installation instructions, usage examples, and new features.
- Enhanced BioServices documentation to clarify API requirements and usage patterns, including email handling for NCBI services.
- Updated Cirq skill documentation to reflect the latest version and installation requirements for Google Quantum AI integration.
- Updated the Medchem library to version 2.0.5, improving drug-likeness rules, structural alerts, and filtering capabilities.
- Enhanced the ESM toolkit documentation, clarifying usage for protein language models and API integration.
- Improved installation instructions and added support for new features, including cloud-based inference and environment variable handling for API keys.
- Streamlined filtering scripts for better usability and performance, including batch processing and error handling improvements.
- Added references for new model versions and updated examples for clarity.
- Updated Biopython version to 1.87, reflecting new requirements for Python 3.10+ and changes in API key handling for NCBI access.
- Clarified installation instructions and deprecated features in Biopython, including the removal of certain modules.
- Enhanced Datamol documentation to specify version compatibility, remote file support, and updated fingerprint generation methods.
- Improved examples and best practices for using Datamol with cloud storage and third-party libraries.
- Added new entries to .gitignore for 'research/', '.agents/', and 'skills-lock.json'.
- Updated version of 'scientific-agent-skills' from 2.37.1 to 2.40.0, reflecting recent changes and improvements.
- Enhanced documentation in SKILL.md and related files for better clarity on usage and installation requirements.
Move SIMBAD content from standalone simbad-database skill into the
consolidated database-lookup skill per reviewer feedback. Enhance
database-lookup/references/simbad.md with detailed script format codes,
full TAP table schemas, ADQL spatial query patterns, cross-matching
examples, bibliography queries, and input sanitization guidance.
Remove the standalone simbad-database directory (SKILL.md, api_reference,
adql_reference, simbad_client.py) since all content is now covered by
database-lookup.
Made-with: Cursor
- Add _sanitize_object_name() to validate script interface inputs (block newlines, quotes, semicolons)
- Add _sanitize_adql_string() to escape quotes and block injection chars in ADQL queries
- Use sanitizers in query_object, query_identifiers, get_all_identifiers
- Add best practice #11 in SKILL.md: validate user input before querying
Made-with: Cursor
Add a new skill for querying the SIMBAD database (CDS Strasbourg),
covering object name resolution, coordinate cone searches, identifier
pattern matching, and advanced ADQL/TAP queries. Includes a Python
client script, script-interface API reference, and ADQL/table schema
reference.
Made-with: Cursor
The Elsevier section of journals_formatting.md referenced
assets/journals/elsevier_article.tex but no such file existed.
Add the upstream elsarticle distribution files (CTAN, LPPL):
- elsarticle-template-{num,num-names,harv}.tex
- elsarticle-{num,num-names,harv}.bst
Rewrite the Elsevier section to:
- map each template to its natbib option and matching .bst
- document the elsarticle-specific frontmatter macros
- list common documentclass options (preprint / review / 1p / 3p / 5p)
- point to elsdoc.pdf for the full class reference
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The description field after PR #149 contained an unquoted 'Note:' clause,
which YAML parsers interpret as a mapping-key separator and reject. Wrapping
the value in single quotes restores parseability while preserving the
PARALLEL_API_KEY / OPENROUTER_API_KEY disclosure intact.
Fixes#159
