2.2K skills · sorted by relevance
Review Python code for common mistakes in error handling, asynchronous operations, type annotations, and tests. Use a practical checklist and correction examples to guide code reviews, investigate bugs, and improve existing implementations.
Split multi-step coding, research, or testing into focused assignments for sub-agents, which are agents handling specific parts of a task. The parent agent retains design decisions, integration, and final verification while independent work runs in parallel.
Retrieve past operational experience from OpenViking before or during coding, configuration, deployment, or recovery work. It reads relevant records, checks whether their assumptions match the current environment, and applies suitable procedures and checks.
Investigate LLM application failures with Phoenix CLI by reviewing execution records, recording observations, and grouping mistakes into counted categories. Inspect datasets and experiments to identify which failure types deserve evaluation tests first.
Create a copilot-instructions.md file from an existing codebase. Identify technology versions, architecture, naming conventions, and testing practices, then turn observed patterns into project-specific guidance for GitHub Copilot without introducing unsupported assumptions.
Generate a customizable AI prompt for finding representative examples in an existing codebase. Configure analysis depth, categories, and snippets so GitHub Copilot can produce exemplars.md with real file references and explanations of coding conventions.
Design, refine, or review web and iOS interfaces using the product’s existing components and visual style. Consult relevant screen examples when useful, then inspect the result for layout problems, inaccessible controls, and broken interactions.
Generate coding agent instruction files from a project's actual code using AgentRC. Capture detected languages, frameworks, and conventions in one shared file or split rules by topic and project area, with a preview before writing.
Examine AI product reliability, learning methods, and model limitations through frameworks drawn from Andrej Karpathy’s public statements. Assess the gap between demonstrations and real use, uneven capabilities, and where people should retain control.
Unwrap layers of obfuscated PowerShell malware to recover hidden commands, additional malicious code, and server addresses. Combines script structure analysis with execution tracing and decoding tools to support incident investigation and detection work.
Compose compact, structured prompts for coding, debugging, review, and research with Codex or GPT-5.4. This internal skill defines task scope, completion criteria, output format, verification steps, and evidence requirements using consistent XML tags.
Turn a project idea into a detailed system plan before coding. Organize architecture, data flows, dependencies, and file-level tasks in a linked Markdown workspace, with research references, consistency checks, and saved state for resuming later.
Guides large language model reinforcement learning with miles, covering environment setup, low-precision training, consistency between training and inference, and speculative decoding, where a smaller model proposes tokens for a larger model to verify.
Provides prompt templates for coding, writing, analysis, and creative work. Helps define the AI's role, specify the desired output, add examples and constraints, and refine requests using a practical checklist.
Set up LLM serving with structured responses, including JSON and outputs restricted to specified patterns. Covers reusing shared prompt calculations, tool calling, and multi-turn conversations through SGLang, with examples for server setup and application integration.
Configure Claude Code and improve how you assign coding tasks. This guide provides a CLAUDE.md template, practical prompting examples, debugging approaches using logs, and a workflow for making and checking small changes.
Extend the amount of text a language model can process at once. Compares RoPE, YaRN, ALiBi, and position interpolation, with implementation examples, additional training guidance, and memory optimization approaches for long document processing.
Combine capabilities from multiple fine-tuned LLMs without retraining using mergekit. Choose a merging method, configure how models are blended, and evaluate the resulting model on tasks such as math, code generation, and conversation.
Control interactive command-line programs through tmux by sending keystrokes, reading terminal output, and waiting for prompts. Use separate sessions to operate Python consoles or run multiple coding agents while checking their progress and results.
Design instructions for agents that write, fix, review, or refactor code. Define behavior based on file access, select suitable tools, and set clear verification and completion rules for repository changes or copyable code output.
Run commands and coding agents on existing cmux Cloud machines, monitor work that continues after disconnection, and present results through workspaces or service URLs. The skill guides machine selection, background execution, and result verification.
Review game code for architecture, coding standards, testability, and performance concerns. Get findings tied to specific code, checks against recorded design decisions, and a report separating required fixes from optional improvements.
Build desktop applications with GPUI Kit, a Rust framework. Guides setup, component selection, state management, themes, and background work, and explains UI integration tests that simulate input and check component behavior.
Design infrastructure for agents that run in isolated remote environments. Plan faster session startup, saved work states, shared sessions, authentication, and delivery of code changes and other results back to users.