Skill UI

Prepares Claude with an ultra-granular, line-by-line code-analysis mode that maps invariants, assumptions, and flows before any vulnerability hunting, keeping context stable and reducing hallucinations during audits.

Prepares Claude with an ultra-granular, line-by-line code-analysis mode that maps invariants, assumptions, and flows before any vulnerability hunting, keeping context stable and reducing hallucinations during audits.

Identifies predictable context degradation patterns in large language models to diagnose failures and design resilient agents; covers lost-in-middle, poisoning, distraction, and confusion with mitigation tactics for long or complex conversations.

Explains what makes up a language model’s context state and how engineers should structure prompts, tools, retrieval, history, and outputs to guide agent behavior and manage token budgets.

Guides how to stretch limited LLM context windows by compressing, masking, caching, and partitioning information so agents process longer conversations or documents without enlarging models; includes when to trigger optimizations and concrete patterns for summaries, masking, KV-cache reuse, and budget tracking.

A comprehensive framework for building persistent memory systems within Large Language Models (LLMs). It supports multi-faceted memory management, including short-term context storage, long-term knowledge retrieval, and entity-based fact updates. This skill is crucial for developing sophisticated AI assistants that maintain coherence, recall user history over extended periods, and provide context-aware interactions by managing memory consolidation and retrieval relevance.

Cursor AI Debug Toolkit helps identify root causes of incorrect Cursor AI suggestions, inspect completion/chat context, and use dev tools/logs to tweak settings, re-test models, and document fixes for future issues.

Extends transformer context windows using RoPE, YaRN, ALiBi, and position interpolation so long documents (32k-128k+ tokens) can be processed, pre-trained models lengthened, and efficient positional encodings or extrapolation strategies deployed.

Guides designers of language-model multi-agent systems through supervisor, swarm, and hierarchical patterns while stressing context isolation, coordination protocols, and token economics for complex research or tool-intensive workflows.

A concise reference for spotting and avoiding Perplexity Sonar anti-patterns during reviews, onboarding, or audits, covering citation parsing, model selection, recency handling, and context limits so integrations stay accurate and cost-effective.

Converts RDF context into belief-desire-intention structures so agents can track perceptions, deliberate plans, and emit explainable RDF, supporting BDI frameworks or LLMs needing structured cognition.

Deploys AI and NLP-powered detection to spot business email compromise by analyzing writing style, behavioral context, and intent, then integrates baselines, transformer models, and response workflows to surface impersonation attempts that rule-based filters miss.