AI & ML

tonone-ai Skill

Build automated regression suites — golden sets, threshold alerting, CI integration for model changes. Use when asked to "catch model regressions in CI", "build a golden set", or "automate eval regression".

tryboy869 Bundle

[Applies to: **/*.py] Definitive guidelines for building robust, maintainable, and scalable multi-agent systems using CrewAI, focusing on modularity, clear role definition, and disciplined LLM configuration.

tryboy869 Bundle

[Applies to: **/*.py] This guide outlines definitive best practices for using the gensim library, focusing on reproducibility, efficient corpus construction, robust model training, and maintainable code for NLP topic modeling tasks.

outlinedriven-odin-claude-plugin Bundle

Use when writing a Linux char driver: file_operations, cdev, copy_to_user, ioctl commands, device memory mmap, or poll. Not for probe and remove lifecycle: use platform-device-model.

tonone-ai Skill

Design production prompts — system prompt architecture, instruction clarity, few-shot selection. Use when asked to "write a system prompt", "design a production prompt", or "pick few-shot examples for this task".

tonone-ai Skill

Compare two or more models statistically — significance testing and error analysis. Use when asked "which model is better", "is this improvement significant", or "compare model performance".

tonone-ai Skill

Produce a threat model — assets, ranked threats, mitigations, accepted risks. Use when asked to "threat model this", "what could go wrong security-wise", "map our attack surface", or before designing any security-sensitive feature.

outlinedriven-odin-claude-plugin Bundle

Use when asked to review learning progress after a milestone. Not for engineering retrospectives, use engineering-retrospective. For agent-environment ones, use agent-environment-retrospective.

outlinedriven-odin-claude-plugin Bundle

Use when implementing or debugging platform_driver probe/remove, sysfs attributes, device properties, or deferred probe on Linux. Not for device tree syntax: use device-tree.

outlinedriven-odin-claude-plugin Bundle

Use when asked to review GitHub Actions for exploitable vulnerabilities, including prompt injection through Claude Code Action, Gemini CLI, or OpenAI Codex. Read-only. Not for general security review.

tonone-ai Skill

Build prompt versioning systems — storage, A/B testing, regression tracking, rollback. Use when asked to "version our prompts", "A/B test prompts", or "roll back a prompt".

tonone-ai Skill

Maintain per-repo and cross-repo changelogs — append structured entries after agent work. Use when asked to "log this change", "update changelog", "what changed", "change history".

outlinedriven-odin-claude-plugin Bundle

Use when a protocol, concurrent algorithm, or design needs a model-checked TLA+ or Alloy spec, or a TLC or Apalache trace needs reading. Not for choosing when to model: use validation-first-driven.

tonone-ai Skill

Design cost reduction strategies — model tiering, prompt compression, caching, batch inference. Use when asked to "reduce our AI costs", "set up model tiering", or "cut LLM spend".

tonone-ai Skill

Design and implement an AI feature integration — model selection, architecture pattern, system prompt, data flow, error handling, cost estimate. Use when asked to "add AI to this", "LLM integration", "add Claude/GPT", or "AI-powered feature".

tonone-ai Skill

Instrument LLM calls with tracing — span structure, token counts, latency, model metadata. Use when asked to "add tracing to our LLM calls", "log prompts and completions", or "instrument token counts".

abelrguezr Bundle

Security skill for understanding, detecting, and defending against prompt injection attacks on AI/LLM systems. Use this skill whenever the user asks about AI security, prompt injection vulnerabilities, LLM attack vectors, jailbreak techniques, or needs help securing AI applications against malicious prompts. Also use when reviewing AI system designs, conducting security assessments of LLM integrations, or creating security documentation for AI systems.

abelrguezr Bundle

Security skill for understanding and testing RCE vulnerabilities in AI/ML model loading. Use this skill whenever the user mentions machine learning models, model deserialization, PyTorch, TensorFlow, Keras, ONNX, or any ML framework loading. Also trigger when discussing model security, pickle vulnerabilities, CVE-2024-12029, CVE-2025-23298, or any AI/ML security audit. This skill helps create educational test payloads, audit vulnerable code, and implement mitigations for model loading RCE attacks.

abelrguezr Bundle

Security auditing and hardening for Model Context Protocol (MCP) servers. Use this skill whenever the user mentions MCP servers, Model Context Protocol, AI agent security, tool poisoning, prompt injection in MCP, Cursor IDE vulnerabilities, Flowise MCP, or any MCP-related CVEs. Also trigger when users want to secure AI agent integrations, audit MCP configurations, or understand MCP attack vectors. Make sure to use this skill for any MCP security questions, even if the user doesn't explicitly mention "security" or "audit".

abelrguezr Bundle

Help users understand and implement deep learning concepts including neural networks, CNNs, RNNs, LLMs, and diffusion models. Use this skill whenever the user asks about deep learning architectures, wants to build neural networks in PyTorch, needs help with training loops, or wants to understand concepts like backpropagation, activation functions, attention mechanisms, or generative models. Make sure to use this skill for any deep learning related questions, code reviews, architecture design, or implementation help.

abelrguezr Bundle

How to assess and document AI security risks using industry frameworks. Use this skill whenever the user mentions AI security, ML vulnerabilities, model risks, LLM security, adversarial attacks, data poisoning, prompt injection, or needs to evaluate AI system safety. Trigger for any request about AI threat modeling, security audits, risk documentation, or compliance with AI security standards.

outlinedriven-odin-claude-plugin Bundle

Use when a human explicitly asks for a full repository agent-compatibility pass returning a scored report with prioritized fixes. Not for tasks that require source or remote-system changes.

outlinedriven-odin-claude-plugin Bundle

Use when C or C++ code needs memory-safety or undefined-behavior guarantees proved with CBMC, or ACSL contracts checked with Frama-C Eva or WP. Not for choosing the proof policy: use proof-driven.

abelrguezr Bundle

How to test for clickjacking vulnerabilities in web applications. Use this skill whenever the user mentions clickjacking, UI redressing, iframe attacks, frame-busting, X-Frame-Options, CSP frame-ancestors, or wants to test if a web page can be embedded in malicious iframes. Also use when testing for doubleclickjacking, SVG filter attacks, or browser extension clickjacking. Make sure to use this skill for any web security assessment involving iframe embedding, form manipulation, or UI overlay attacks.

abelrguezr Bundle

How to tokenize text for LLMs and NLP models. Use this skill whenever the user needs to convert text into token IDs, understand tokenization methods (BPE, WordPiece, Unigram), work with vocabularies, or implement tokenization for machine learning. Make sure to use this skill when users mention tokenizing, token IDs, vocabulary creation, BPE, WordPiece, or any text preprocessing for ML models.

abelrguezr Bundle

How to prepare and sample text data for training large language models. Use this skill whenever the user mentions data preparation, tokenization, sliding windows, sequence generation, training data, LLM datasets, or needs to create input/target pairs for model training. This includes tasks like chunking text, creating dataloaders, applying sampling strategies, or optimizing training data quality.

abelrguezr Bundle

Implement and analyze reinforcement learning algorithms (Q-Learning, SARSA, PPO) and detect security vulnerabilities like backdoor attacks and reward poisoning. Use this skill whenever you need to build RL agents, understand RL algorithms, implement training loops, or audit RL systems for security issues. Make sure to use this skill when the user mentions reinforcement learning, Q-learning, SARSA, RL training, agent training, policy learning, or any ML system that learns from rewards and environment interaction.

abelrguezr Bundle

Create and work with token embeddings for LLMs. Use this skill whenever you need to understand token embeddings, create embedding layers in PyTorch, add positional embeddings (absolute, relative, or RoPE), or debug embedding-related issues in your language model. This skill covers vocabulary setup, embedding initialization, positional encoding strategies, and context window extension techniques. Make sure to use this skill when working with any LLM architecture, training pipelines, or when you need to convert tokens to numerical vectors.

abelrguezr Bundle

Build and understand LLM architecture from scratch. Use this skill whenever the user needs to create GPT models, implement transformer components (attention, feedforward, layer norm), calculate model parameters, or generate text with a trained model. Trigger for any request about LLM architecture, transformer blocks, GPT implementation, token embeddings, positional embeddings, or building neural networks for language modeling.

abelrguezr Bundle

Guide for building and training large language models from scratch. Use this skill whenever the user wants to understand LLM training concepts, implement tokenization, data sampling, embeddings, attention mechanisms, model architecture, pre-training, or fine-tuning workflows. Trigger on mentions of LLM training, building models from scratch, tokenization, embeddings, attention, pre-training, fine-tuning, LoRA, or any LLM development task.

abelrguezr Bundle

Prepare and evaluate machine learning data. Use this skill whenever the user needs to clean, transform, or split datasets for ML training, or evaluate model performance with metrics like accuracy, precision, recall, F1, ROC-AUC, MAE, or confusion matrices. Trigger for any data preprocessing task, feature engineering, handling missing values, encoding categorical variables, normalization, or model evaluation requests.

abelrguezr Bundle

Explain and teach Large Language Model fundamentals including pretraining, model architecture, PyTorch tensors, automatic differentiation, and backpropagation. Use this skill whenever the user asks about LLM concepts, neural network training, PyTorch operations, gradient computation, or wants to understand how LLMs work internally. Trigger on questions about model parameters, context length, embedding dimensions, tensor operations, autograd, or backpropagation.

abelrguezr Bundle

How to implement and understand attention mechanisms in neural networks and LLMs. Use this skill whenever the user needs to build self-attention layers, causal attention, multi-head attention, or understand how attention weights are calculated. Trigger this skill for any task involving attention scores, Q/K/V matrices, attention masking, or transformer architecture components.

abelrguezr Bundle

How to research and understand Python security vulnerabilities including sandbox escapes, deserialization attacks, and Pyscript exploitation. Use this skill whenever the user mentions Python security, sandbox bypass, deserialization vulnerabilities, Pyscript hacking, Keras model attacks, or needs to understand Python-based attack vectors for security research, penetration testing, or defensive analysis.

abelrguezr Bundle

How to train LLMs from scratch using PyTorch, including model architecture setup, data preparation, training loops, loss monitoring, and model saving/loading. Use this skill whenever the user wants to train a language model from scratch, understand pre-training workflows, set up GPT architectures, configure training parameters, monitor loss/perplexity, or load/save model checkpoints. Make sure to use this skill when users mention training LLMs, pre-training, model checkpoints, GPT architectures, training loops, or want to build language models from the ground up.

abelrguezr Bundle

How to fine-tune a pre-trained LLM (like GPT2) for text classification tasks. Use this skill whenever the user wants to adapt a language model for classification (spam detection, sentiment analysis, topic categorization, intent classification, or any binary/multi-class text classification). Trigger this skill when users mention fine-tuning, classification, adapting models, or need to convert a generative model into a classifier.