Claude Skills
The open format is called Agent Skills and works in Claude Code, Codex, Cursor and other agents — most people know it as Claude Skills.
Every Agent Skill we could find on GitHub, deduplicated by content. 79 566 files from 1 758 authors, of which 61 913 are unique — the rest is the same skill repackaged into someone else's repository. For each one: what it weighs in tokens, whether it ships runnable scripts, and which MCP servers it needs.
Use this skill to identify, extract, and populate specific fields within a PDF form using the pdftk or pikepdf toolsets.
Evaluating clustering performance using greedy matching of predicted centroids against ground truth points.
Efficiently executing grid search over massive parameter spaces using Python multiprocessing.
Identifying Pareto-optimal points in a multi-objective optimization space.
Logic to parse nested security database objects into a scalar CVSS score for the final CSV report.
Analyze /root/package-lock.json to identify HIGH and CRITICAL vulnerabilities using automated security audit tools and database lookups.
Construct the final CSV output file ensuring strict column compliance with the security audit requirements.
Compute the distance from each Pacific plate earthquake to the unified plate boundary.
Format and save the identified earthquake data as a JSON file.
Load earthquake GeoJSON and plate boundary datasets, and ensure consistent projection for distance analysis.
Use this to structure the final output into the required JSON schema, calculating precise token usage based on the retrieved context.
Use this to search, filter, and navigate enterprise artifacts (/root/DATA) to answer specific questions, ensuring cross-product isolation and multi-hop link traversal.
Use this to ingest the task file and orchestrate the multi-step retrieval and formatting workflow.
Compares holding changes between two quarters for a specific fund to determine investment increases.
Searches all funds within a specific quarter to rank holdings for a target security by CUSIP.
Ensures the generated output strictly adheres to the required JSON schema.
Extracts AUM and total holdings count for a specific fund using its accession number and the quarter folder path.
Applying patches to the Druid source code and performing a targeted Maven build.
Neutralizing malicious logic by modifying class deserialization behaviors.
Generating and saving git-compatible patch files for Apache Druid source code.
Used to transform raw GitHub CLI output into the required reporting structure, ensuring data integrity and correct typing.
Used to query, filter, and extract metadata from GitHub PRs and Issues using the `gh` command line tool with appropriate search qualifiers and pagination limits.
Synchronizes the Python environment with specific dependency versions defined in project files to ensure numerical reproducibility.
Implements the SimPO loss function using stable log-probability operations in the specified trainer file.
Creates an orchestration script to run the immutable unit test and capture the output loss values.
Strategies for programmatically modifying .docx files, handling placeholders, and performing conditional text removal.
How to read and extract specific values from employee_data.json to map them to template placeholders.
Robustly extracts text from PDF, DOCX, and PPTX files while maintaining structural integrity for multi-column layouts and nested elements.
Safely moves files into target directories while preventing collisions, handling path logic, and maintaining state.
Maps documents to specific categories using a combination of keyword-based semantic matching and categorical hierarchy.
Aligning the implementation with provided Test Specifications.
Translating Python dynamic structures into type-safe Scala ADTs and case classes.
Implementing tokenization logic using functional patterns and standard library collections.
Extracts meeting requirements from /root/test_input.json, handling variable constraints and field parsing.
Resolves scheduling conflicts between email requests and calendar availability, prioritizing meeting requests over blue (flexible) blocks.
Parses /root/calendar.pdf to extract appointment timings, color-coded blocks, and time-axis mappings using pdfplumber.
Creates the directory structure, copies data, and writes the specific JS/CSS/HTML files required for the web application.
Defines the layout, chart aesthetics, and table styling in style.css.
Implements the D3.js force simulation, tooltips, and data interaction logic in visualization.js.
Performs efficient parameter optimization using a heuristic search or scipy.optimize to satisfy RMSE constraints within fixed parameter ranges.
Calculates RMSE between simulated and observed data using exact datetime matching and integer depth binning based on simulation start date 2009-01-01.
Executes the GLM simulation and retrieves lake-specific depth data to ensure accurate mapping between model output and field observations.
Safely updates specific calibration parameters in the glm3.nml file using f90nml, maintaining correct namelist group mapping and file integrity.
Instructions for querying the provided datasets to ensure constraints regarding pet-friendly lodging, specific cuisines, and valid locations are met.
Ensures the final output strictly adheres to the requested JSON structure and field requirements.
Determines the driving route from Minneapolis to Ohio, segments the 7-day travel period into manageable travel and stay durations, and ensures no flights are used.
Aggregate object counts into a structured CSV file.
Convert images to grayscale in-place using OpenCV.
Count objects in an image using the command-line interface of the object detection script.
Use ffmpeg to extract keyframes from a video file into a target directory.
Use INDEX combined with MATCH to retrieve data from a source table based on two criteria (row and column headers).
Perform descriptive statistical analysis while ensuring results are displayed as percentages rounded to one decimal place.
Calculate the weighted mean by multiplying Net Exports % by the corresponding GDP values and dividing by the total GDP.
Guidelines for mapping modern Mandarin pronunciation (Pinyin) to traditional Ping (Level) and Ze (Oblique) poetic tones, and identifying modern rhyme families.
Structural, syntactical, and tonal rules for composing a seven-character regulated verse (七言律诗).
Inspecting and filling interactive PDF forms (AcroForms) using the pypdf library. Crucial for automating the population of government or legal PDF documents.
Efficiently computing DBSCAN with a custom weighted (anisotropic) distance metric using coordinate scaling, avoiding the overhead of slow custom distance functions.
How to perform spatial filtering (point-in-polygon), combine line boundaries, and accurately calculate metric distances between points and boundaries using GeoPandas.
Claude Skills — questions
Answers built from the skills we actually parsed.