Skill UI

Conduct authorized security assessments for LLM applications and AI agents. Covers prompt injection, tool abuse, RAG exposure, memory poisoning, and system prompt extraction following OWASP LLM/ASI Top 10 guidelines. Includes red-teaming, vulnerability mapping, and compliance engineering for AI systems to ensure safety and robustness.

Luna is an automated code reviewer focused on correctness, security, and reliability. She evaluates code against blueprints and checklists, identifying critical vulnerabilities like injection flaws, authentication bypasses, and race conditions. Luna acts as a quality gate, ensuring no unresolved high-risk issues proceed to QA or deployment. She excludes style preferences, focusing only on measurable risks to maintainability and safety.

Build multichannel enterprise agents for Microsoft 365, Teams, and Copilot Studio using the Microsoft.Agents SDK for .NET, with ASP.NET Core hosting, AgentApplication routing, and MSAL authentication.

Build enterprise agents for Microsoft 365, Teams, and Copilot Studio using the Microsoft Agents SDK for Python. Supports aiohttp hosting, AgentApplication routing, streaming responses, and MSAL-based authentication.

Red-team playbook for Microsoft 365 and Entra ID: tenant discovery, AADSTS error-code triage, Smart Lockout math for safe credential spraying, and user-enumeration vectors. Authorized engagements only.

Authorized macOS and Mach-O reverse engineering: inspect code signatures, entitlements and Hardened Runtime, recover Objective-C/Swift structures, and triage Apple-platform malware with otool, Hopper, Ghidra, Frida and lldb.

Analyze suspected malware through static, dynamic, and behavioral techniques: IOC extraction, YARA/Sigma rule authoring, sandbox orchestration, and anti-analysis detection.

Audit multi-role marketplace authorization across roles, resource ownership, tenant boundaries, and order-state transitions. Use this skill when access rules require evidence instead of UI assumptions, covering policy matrices, identity tracing, and field-level security checks.

Comprehensive hardening guide for Model Context Protocol (MCP) servers, covering transport encryption, mTLS, authentication, authorization, input validation, and audit logging to secure AI agent integrations.

This skill enables security auditing of meme coins and tokens on EVM and Solana blockchains. It identifies critical vulnerabilities like hidden mint functions, honeypot transfer restrictions, fee manipulation, and liquidity pool drains. Users must obtain explicit authorization before scanning targets to ensure ethical use.

Leverage the Metasploit Framework for authorized penetration testing, from module search and exploit configuration to payload generation, auxiliary scanning, and post-exploitation. Use only with explicit written permission.

Microsoft Entra Authentication Events SDK for .NET. Azure Functions triggers that handle custom authentication extensions, including token claim enrichment, attribute collection customization and validation, and custom OTP delivery.