Skill UI

Govern evidence-backed canonical project state across sessions, branches, reviews, and research cycles. Maintains authoritative status, decisions, constraints, and negative evidence so any agent resuming work can verify claims, avoid repeated failures, and know the next step.

Protect AI systems from prompt injection attacks by implementing input sanitization, tool permissioning, and output validation. Essential for securing LLM applications, RAG pipelines, and agentic workflows against indirect and direct attacks.

A governance skill designed to secure AI agent tool calls using advanced authorization policies. It implements Cedar policies to control which tools agents can access and under what conditions. Key features include running in 'Shadow mode' for safe observation, 'Enforce mode' for strict blocking, and generating cryptographically signed Ed25519 receipts to provide an immutable, tamper-evident audit trail for every policy decision.

PydanticAI is a robust Python framework that integrates Pydantic's type-checking and validation capabilities into Large Language Models (LLMs). It allows developers to build production-ready AI agents that guarantee structured, validated outputs, moving beyond simple raw strings. Key features include supported tool calling, dependency injection for testability, and multi-model compatibility across major providers like OpenAI, Anthropic, and Gemini. Use this when reliability and correct data structure are paramount.

Guides an AI agent to give non-judgmental, evidence-informed quit-smoking support: quit planning, craving check-ins, slip debriefs, optional consent-based logging, and escalation to clinicians or quitlines. Not medical care.

RecallMax dramatically enhances AI agent memory capabilities by cleanly injecting massive external contexts (500K-1M tokens) without hallucination drift. It automatically summarizes lengthy conversations, preserving tone, intent, and key facts. Furthermore, it compresses multi-turn interaction histories into high-density token sequences, ensuring agents maintain perfect recall across complex and long-running sessions.

This skill implements advanced 'Gatekeeper' logic to optimize AI agent performance in long-running workflows. It prevents context window bloat by sharding metadata, allocating token budgets, and enforcing Atomic Precision to deliver purely functional, filler-free responses. Ideal for complex coding, multi-step reasoning, and large document analysis where conciseness is critical.

This skill reviews git diffs or specific file scopes for code reuse, quality, efficiency, clarity, and standards compliance. It optionally applies safe, behavior-preserving fixes. It uses parallel read-only sub-agents to inspect code while the main agent applies changes. Suitable for auditing changes, simplifying code, or refactoring.

This skill enables parallel read-only multi-agent review of git diffs or specific file scopes. It identifies behavioral regressions, security risks, performance issues, and coverage gaps using four specialized sub-agents. The main agent synthesizes findings to filter and summarize critical issues without modifying code. Ideal for ensuring code quality and safety before merging changes.

Rex translates vague user intent into precise, unambiguous specifications. It extracts core problems, defines audiences, identifies edge cases, and produces structured reports with user stories, constraints, and open questions. Ideal for clarifying requirements before development.

Simulates Sam Altman, CEO of OpenAI and former Y Combinator president, to provide insights on startups, AGI, fundraising, and the future of AI. Use for advice in the style of Sam Altman.

This skill enables AI agents to search, import, and analyze video content using Memories.ai's Large Visual Memory Model. It provides persistent video intelligence, allowing videos to be indexed once and queried repeatedly. Supports social media import from TikTok, YouTube, and Instagram, meeting summarization, knowledge base building, and cross-video Q&A via Memory Augmented Generation. Ideal for analyzing video URLs, researching social media trends, and summarizing recordings.