Skill UI

Manages the complete lifecycle of GitHub issues using a structured, label-based state machine. This skill assists maintainers in triaging incoming reports, determining if an issue is a bug or enhancement, and moving it through defined states (e.g., needs info, ready for agent, wontfix). It ensures systematic workflow management for development projects.

This skill orchestrates the entire development lifecycle for approved software tasks. It operates in autonomous or manual modes, rigorously applying Test-Driven Development (TDD). It automatically gathers all necessary context (specs, requirements, design, and historical learnings), validates the repository state, and dispatches specialized sub-agents to implement tasks one by one, ensuring high code quality, adherence to design, and process traceability.

This tool systematically converts a project roadmap into actionable specifications. It first analyzes features by grouping them into dependency waves. It then automatically dispatches sub-agents to generate complete feature specs (requirements, design, tasks, and JSON) for each wave. The core value lies in its comprehensive cross-spec review, which verifies data model consistency, interface alignment, and architectural integrity across the entire project, ensuring a cohesive final product.

A comprehensive index of 46 production-grade test automation skills, covering E2E, unit, mobile, BDD, and cloud testing across 15+ languages. This skill equips AI coding assistants to act as senior QA architects, generating structured, real-world test code for major frameworks like Playwright, Selenium, Cypress, and Jest. It supports project setup, CI/CD integration, and cross-platform debugging.

This guide teaches how to build robust, production-grade ReAct agents using LangGraph 1.0's `create_react_agent`. It specifically addresses common agent failure modes, such as preventing infinite loops (recursion caps) and ensuring proper error propagation. Learn to define typed tools using Pydantic, manage state with checkpointers, and set precise recursion limits. Perfect for users migrating from legacy AgentExecutor or building reliable, multi-step tool-calling workflows.

A comprehensive guide to ensuring robust state management for LangGraph agents. This skill addresses common checkpointing pitfalls (like missing thread_id or non-JSON serializable state) and details best practices for implementing reliable state persistence using MemorySaver or PostgresSaver. It covers thread discipline, schema migration, and time-travel debugging for production-ready applications.

This skill defines canonical, ordered middleware patterns for LangChain and LangGraph agents. It addresses critical production-level LLM deployment issues, including PII leakage across tenants, token budget overconsumption, prompt injection defense (Guardrails), and cache key corruption. Use this to build secure, scalable, and cost-aware AI applications.

This framework implements the OWASP Top 10 standards, specifically tailored for Agentic Applications. It provides a standardized methodology to classify and map diverse security findings—such as prompt injection, tool misuse, and privilege abuse—to ten core risk categories (ASI01-ASI10). It is essential for security teams to systematically assess, prioritize, and remediate vulnerabilities inherent in complex AI agent systems.

This internal agent skill automates the complex process of determining the most likely owner or team for a bug cluster. It follows a strict 6-level precedence model, starting with cached overrides, and sequentially querying sources like Service Owners, Oncall status, CODEOWNERS, recent assignees, and committers. The process includes staleness detection and confidence scoring, providing highly accurate and auditable routing recommendations for internal development triage.

This skill facilitates an interactive Quality Assurance (QA) session. Users can conversationally describe encountered bugs or issues. The agent clarifies the problem, explores the codebase for context and domain language, and then meticulously files durable GitHub issues. It supports both single-issue reports and breaking down complex problems into multiple, independently verifiable tickets, ensuring developers receive actionable, high-quality reports written from the user's perspective.

This internal agent skill defines a systematic, multi-step procedure for scanning GitHub repositories to gather comprehensive evidence supporting or explaining bug clusters. The process covers searching open issues, inspecting recent commits, analyzing code paths, and checking deployment history. It assigns evidence confidence tiers (Exact, Strong, Moderate, Weak) to provide a deep, actionable understanding of potential software defects.

Shipwright is an autonomous, multi-phase agent designed to convert simple, plain-English requirements into fully functional, tested, and deployment-ready applications. It supports modern full-stack frameworks like Next.js/Supabase, SvelteKit, and Astro. Use this tool when you need to scaffold a complete project, add complex features, or generate an entire SaaS dashboard merely by describing it.