Skill UI

This skill enables AI to adapt its operational behavior based on specific task types such as brainstorming, coding, debugging, reviewing, teaching, and deployment. It optimizes performance by switching between distinct modes like BRAINSTORM, IMPLEMENT, and DEBUG, ensuring appropriate communication styles and output formats for each scenario.

This skill prevents AI coding agents from entering ineffective repair loops by establishing a bounded repair contract. It uses stable failure fingerprints, a three-attempt budget, and real-path verification to ensure progress. When fixes do not change observable states, it mandates shifting root-cause strategies rather than retrying. Includes rollback controls and negative testing to validate actual system behavior over agent claims.

This skill provides universal coding standards, best practices, and design patterns for TypeScript, JavaScript, React, and Node.js development projects. It ensures consistency and quality across codebases while referencing detailed guides for execution.

A comprehensive utility for managing complex cmux topologies, including workspaces, panes, and surfaces. This tool is essential for inspecting, controlling, and monitoring multiple AI coding agents running in parallel within a macOS terminal environment. It supports advanced functionalities such as browser automation and real-time status notifications, optimizing developer workflows.

Runs Codex CLI non-interactively for code analysis and implementation tasks. Delegates repository work to Codex and obtains coding-agent results via JSONL output. Supports sandbox modes and authentication checks.

Runs the Hermes Agent CLI in one-shot mode (hermes -z) to delegate bounded local coding tasks from Code Mate. Covers availability checks, timeouts, exit codes, authentication limits and YOLO tool-approval safety constraints.

This skill automates Kimi Code execution in non-interactive mode within a bounded repository. It parses event streams, handles authentication securely, and manages sandbox permissions. Use it to delegate specific coding tasks to Kimi Code while ensuring safety constraints like timeouts and directory limits. It prevents interactive UI launches and avoids credential exposure.

Runs MiniMax Code headlessly for repository analysis and coding tasks. Delegates work to MiniMax Code, checks availability, executes tasks with proper permissions, and summarizes results.

Runs Qwen Code headlessly for repository analysis and coding tasks. Delegates work to the AI agent, parses JSON output, and handles authentication errors. Use for automated code review, bug analysis, and refactoring without interactive UI.

This skill simplifies and refines code to enhance clarity, consistency, and maintainability without altering functionality. It applies project-specific best practices to improve readability, reduce complexity, and align with coding standards. Ideal for cleaning up recently modified code or optimizing existing logic for better structure.

This skill enables orchestrating agents to delegate bounded coding tasks to the OpenAI Codex CLI. The orchestrator drafts the brief, runs the command, and maintains full responsibility for reviewing diffs and committing changes. It prioritizes safety by preventing automatic commits, requiring explicit user requests and manual gate verification.

Delegate bounded coding tasks to the Command Code CLI (cmd) while retaining review and landing responsibility. The orchestrator writes a brief, dispatches via a relay script, reviews the diff, and commits. Requires cmd CLI installed and authenticated, Node 22+, and git.