Claude Skills by ECNU-ICALK

Converts Python PEG grammar definitions into a Rust-like Ungrammar format, specifically optimizing for concise top-level structures while maintaining detailed pattern definitions and specific operator placement.

Simulates a Python REPL environment that executes code silently and supports a custom `prompt()` function to access GPT capabilities for generating data structures or text.

Generate Python code using machine learning to predict the probability of specific colors (red, purple, yellow) in a roulette game based on historical data, and calculate the model's accuracy.

Implements a Python-based query executor that converts informal keywords to SQL commands, parses queries preserving string literals, and executes operations on pandas DataFrames.

Create a user-friendly Python GUI using Tkinter that allows users to load a trained Keras model via a file explorer and select a tokenizer from a dropdown menu to perform text inference.

Generates a Python function using PIL to composite a masked overlay onto a background and render auto-fitting text. The text is centered, shadowed, uppercased, uses baseline-based line spacing, and randomly highlights words.

Develop a PyTorch-based simple diffusion neural network to generate 16x16x16 matrices. The implementation must include a custom dataset loader that reads .raw files from a 'dataset/' directory, extracts the text prompt from the filename, and saves generated results to an 'outputs/' directory.

Implement a CNN image classifier in PyTorch with specific architectural constraints (6 conv layers, residual connections), PyTorch-native data splitting, and code-heavy output.

Implements a PyTorch Transformer model with configurable layer dimensions and attention masking, and a training loop that retains the best performing model based on validation loss.

Modifies PyTorch data preparation scripts for RNN/LSTM models to limit the dataset size by dividing it into chunks controlled by a hyperparameter, ensuring the first dimension of input/target tensors fits memory constraints.

Implements PyTorch functions for hard negative mining and triplet loss calculation using cosine similarity, specifically handling scenarios where anchors have multiple positive samples and requiring mask-based operations.

Configure and train a Mixture of Experts (MoE) Transformer model in PyTorch, implementing a custom GELU activation function, learning rate warmup, and comprehensive evaluation metrics (Precision, Recall, F1).

Debugs PyTorch dimension mismatch errors by adding print statements to inspect tensor shapes at key points in the model forward pass and training loop.

Provides a complete end-to-end workflow for text classification using a PyTorch Transformer model. It includes automatic vocabulary generation from raw text, a custom tokenizer implementation, data padding, model training on CPU, and visualization of loss and accuracy metrics.

Generate complete R code for hierarchical Bayesian models using Gibbs/Metropolis sampling, strictly adhering to a user-provided template that includes initialization, sampling, convergence diagnostics (trace/ACF), multi-chain execution, thinning, and chain combination.

Execute a comprehensive 4-step financial analysis in R using dplyr and PortfolioAnalytics: summary statistics, constrained asset selection (Reward/Risk and P/E based), data exploration, and portfolio optimization (GMVP and Tangency) using the ROI solver with specific constraints.

Evaluate API calls on a 1-3 scale based on syntax correctness, parameter optimality, and retrieval likelihood.

Systematically analyze a rational function to determine its domain, intercepts, asymptotes, and behavior across intervals to prepare for graphing.

Implement a React product card component that features a mouse-relative 3D tilt effect and a hover-triggered RGB glowing border using CSS pseudo-elements.

Renders a selected concept with its parent and child relationships in a hierarchical layout using Ant Design components, emphasizing the selected concept.

Implement a full-stack React and Flask solution for creating and updating 'annonce' entries with media. Features include client-side previews, upload progress tracking, and server-side file management (saving new files and deleting old ones during updates).

Develops a Python script to merge housing datasets, perform regression with RandomForestRegressor, create a binary classification target based on median price, and generate specific metrics (MAE, R2, F1, Accuracy) and visualizations (ROC, Confusion Matrix, Density Plots).

Develops, debugs, and adapts Open Shading Language (OSL) shaders specifically for Redshift in Cinema 4D. Ensures compatibility across different geometries (cubes, spheres, planes) by using UV coordinates and adheres to Redshift-specific syntax constraints.

Generates a Cinema 4D Redshift OSL shader for a multi-color pulsing effect, strictly adhering to Redshift metadata syntax and variable scope constraints.

Transforms decompiled C code (e.g., from Ghidra or IDA Pro) into clean, readable C99 code by replacing generic variable names with descriptive ones, standardizing types, and adding explanatory comments.

Refactor imperative three-globe configuration chains into a declarative React component using globe.gl that accepts visualization parameters as props.

Migrates a native HTML select dropdown to the react-select library, applying a black background theme and handling the onChange event signature difference.

Refines acceptance criteria for software tickets by strictly adhering to the problem statement, excluding non-functional requirements (ease of use, error handling, accessibility), and structuring output into Acceptance Criteria and Validation Checkpoints.

Transforms raw user instructions into optimized, single-line prompts suitable for LLMs, acting as a prompt engineer.

General SOP for common requests related to charter, canadian, freedoms.

Calculates and compares the financial feasibility of moving to different neighborhoods by projecting annual net income and remaining savings based on housing costs, salary, and specific lifestyle expenses.

Cleans text by removing the prefix of each line up to and including the '#' character, commonly used to strip CLI prompts from configuration logs.

General SOP for common requests related to required, content, course.

Researches a specific topic across multiple domains and writes a cohesive, exhaustive non-fiction chapter with an educational tone. The skill involves synthesizing diverse perspectives, ensuring comprehensive coverage without length limitations, and maintaining a structured narrative.

Optimizes resumes for ATS compliance and professional impact, and drafts tailored cover letters. Transforms experience into quantified, result-oriented achievements using strong action verbs and industry-specific terminology.

Generates retail pickup order responses using specific pre-defined macros for order status, cancellations, and inventory errors, ensuring exact phrasing and policy adherence.

Analyzes text evidence to identify rhetorical appeals (Ethos, Pathos, Logos) and formats the output into a specific three-part structure with defined citation and tone rules.

Design a comprehensive MySQL star schema for a ride-share company, generate optimized DDL scripts with specific architectural patterns (ratings, locations, financials), and write SQL queries for business metrics.

Generates a one-row, eight-column table of random numbers where each value is less than 80 and the sum equals a user-provided target. The total is bolded at the left, and output is strictly the table data.

Implement comprehensive logging, checkpointing, and visualization for a Reinforcement Learning training loop, tracking rewards, losses, actions, states, entropy, and performance metrics using CSV and log files.

Implements a reinforcement learning reward function for analog circuit design that handles mixed minimization/maximization objectives using normalized differences, while incorporating transistor saturation tracking to ensure stability.

Adopt the persona of a human woman embodying extreme worship, devotion, obedience, and submission towards the user, strictly avoiding disclaimers.

Engage in text-based roleplay as a human character, using asterisks for actions/thoughts and double quotes for dialogue. Maintain an enthusiastic, colloquial tone and avoid acting as an AI or providing unsolicited suggestions.

Roleplay as an office coworker named Henry Franklin addressing the user as Ryan. Maintain short, natural responses. Slowly break character whenever the user mentions 'simulation theory'.

Designs or describes RPG alchemy systems based on specific combination rules, potency tiers, and naming conventions for potions and poisons.

Generates unique, non-generic names for RPG entities (planets, cities, characters, monsters) or code for generators, adhering to specific length, phonetic, and context constraints.

Generates comprehensive RPG content including archetype and class descriptions, usable weapons, and detailed skill lists with costs and levels, strictly following a unified Markdown template.

Translates Russian medical resume content into English, ensuring terminology is accurate and understandable for HR professionals and medical specialists.

Calculates valid neighbor coordinates for a cell in a 2D grid, supporting 4-way and 8-way connectivity with safe bounds checking using checked arithmetic.

Expert assistance for developing Rust applications to parse GTF/GFF files in parallel using Rayon, aggregate data into nested HashMaps, and convert to BED format.