Results for “pre-pr”

Provide graduated assistance from abstract conceptual nudge to concrete procedural step, with reflection required before each escalation. Teaches help-seeking as a skill and prevents direct-answer shortcuts.

Extends effective context capacity through strategic compression, masking, caching, and partitioning techniques.

Automates Azure OpenAI model deployment by checking capacity across regions and deploying to the best available option.

Quantize LLMs to 8/4/3/2/1-bit precision without calibration data, using multiple backends and HuggingFace/vLLM integration.

Adapt a classroom task for specific learner needs while preserving the core learning objective intact. Use when differentiating for SEND, EAL, gifted, ADHD, dyslexia, or anxiety.

Design metacognitive checkpoints that prevent AI-assisted learning from bypassing genuine understanding. Use when students use AI tools and may overestimate their own comprehension.

Trigger: security rules, credential leaks, auth tokens, input sanitization, threat model. Scope: Software security guidelines, threat prevention, secure data handling. Boundary: Excludes client-side web browser vulnerabilities.

Persistent cross-session memory for AI agents. Mem0 stores user preferences, past decisions, domain knowledge, and agent learnings across all sessions, all tools, and all users. Complements planning-with-files (task-level memory) with long-term agent intelligence (CRM + personal knowledge base layer). Use when asked to "remember this", "store preference", "mem0", "long-term memory", "user memory", "agent memory", or when building multi-session agents that need to recall past interactions.

Computes Theil's U (uncertainty coefficient) between predictions and ground truth using the torchmetrics implementation, handling categorical data and NaN strategies.

Tier vocabulary from a text or topic into everyday, academic, and technical categories with teaching priorities. Use when pre-teaching vocabulary or identifying language barriers in a text.

Compact the current conversation into a handoff document for another agent to pick up. Use when user wants to hand off work, summarize progress, or prepare context for a new session.

AI-driven Game Development Studio (BMAD-GDS). Routes game projects through Pre-production, Design, Architecture, Production, and Game Testing phases using 6 specialized agents. Supports Unity, Unreal Engine, Godot, and custom engines.

Generates Python code to batch-translate the English column of tab-separated CSV files while preserving the original Chinese column and output format.

Classifies situations into 256 binary states and maps each to a prescribed action, reporting the pattern, decimal, name, range, and action to execute.

Predicts enhancer activity in DNA sequences using the hosted Genomic Intelligence G0 DeepSTARR model, returning per-window activity scores.

Process whole slide images for digital pathology: detect tissue, extract tiles, and prepare datasets for deep learning pipelines.

Explains machine learning model predictions using SHAP values, covering feature importance, visualization plots, model debugging, bias analysis, and production deployment.

Present pathway options and orchestrate place-based curriculum design from a local place, curriculum requirement, or community issue. Use when place should become a primary text for learning.

Add Adaptive Card response templates to MCP-based API plugins for visual data presentation in Microsoft 365 Copilot.

Train, fine-tune, and deploy LLMs using LitGPT's clean implementations of 20+ architectures like Llama, Gemma, and Phi.

Access AI-ready drug discovery datasets, benchmarks, and molecular oracles from Therapeutics Data Commons for therapeutic machine learning and pharmacological prediction.

Chunks a book into canonical retrieval units with heading-aware structure splitting, recursive token targets, and contextual prefixes for downstream RAG ingestion.

Discovers, finds, audits, repairs, adapts, crafts, runs, debriefs, saves, and prepares repeatable AI-agent loops for publication, treating loops as bounded feedback systems.

Computes the SQuAD metric using torchmetrics, given predictions and ground truth. Use when evaluating question-answering outputs with exact match and F1 scores.

Creates Azure OpenAI model deployments with intelligent intent-based routing, supporting quick presets, full customization, and capacity discovery across regions and projects.

Generates images via OpenAI, Google, DashScope, and Replicate APIs, supporting text prompts, reference images, aspect ratios, and quality presets.

Converts code repositories and document directories into semantically-chunked text files optimized for NotebookLM ingestion, with support for config files and incremental processing.

Train machine learning models end-to-end, covering data loading, preprocessing, architecture selection, training loops, validation, and checkpointing. Use when the user requests model training or provides relevant inputs for this workflow.

Runs a structured clarifying interview for new project requests before building. Instead of writing code, it outputs a fully specified prompt.md for a fresh agent session to execute, preventing expensive mistakes.

[DEPRECATED] Use `create-video` for prompt-based video generation or `avatar-video` for precise avatar/scene control. This legacy skill combines both workflows — the newer focused skills provide clearer guidance.

Serve LLMs and VLMs on NVIDIA Jetson devices using vLLM or SGLang with optimized Docker containers and quantization presets.

Build and evaluate machine learning models using scikit-learn for classification, regression, clustering, dimensionality reduction, and preprocessing.

Formulate testable hypotheses from observations, design experiments, and generate predictions using a structured scientific method framework.

Translates a Chinese article into English Markdown, generates bilingual detailed summaries, and writes the results back to LanceDB while preserving structure.

Reduce high-dimensional data with UMAP for visualization, clustering preprocessing, and supervised or semi-supervised learning, including parameter tuning guidance.

Evaluates how well internal model representations (hidden states) predict token-level information importance in summarization tasks, using NDCG@10 and Spearman's rank correlation.