Shape agent capabilities into your workflow.

Use one CLI to find, install, update, or remove skills across Claude Code, Codex, Cursor, OpenCode, and many other agents. It can also generate a selected skill prompt or launch a supported agent without installing the skill into an agent directory. Sources can come from GitHub, GitLab, generic Git URLs, or local paths, with project or global installation; support for advanced fields still varies by agent.

Deep Agents is an agent harness for long-running, multi-step work, with built-in planning, sub-agents, filesystem access, shell use, and persistence. Each part can be overridden or replaced without forking, and the framework can work with tool-calling frontier, open-weight, or local models.

Kilo Code works in VS Code, JetBrains, and the terminal to generate multi-file changes, autocomplete code, run commands, and automate browser tasks. You can switch models during a task and choose specialized agents for planning, debugging, questions, or review. Its autonomous CLI mode removes permission prompts, so it should be limited to trusted environments.

Qwen Code is an open-source coding agent for the terminal with Auto-Memory, Auto-Skills, subagents, agent teams, and MCP. It supports interactive and headless use, an experimental shared daemon, and providers using OpenAI, Anthropic, Gemini, Qwen, or local-model-compatible APIs. The project also ships IDE, desktop, SDK, and IM integrations; Node.js 22 or newer is required for the npm installation path.

Store task plans, findings, and progress in Markdown files so coding agents can recover after context loss, session clearing, crashes, or handoffs between workers. Optional lifecycle hooks remind the agent to reread and update the plan, while gated modes prevent it from stopping before explicit completion conditions are met. The skill works across many SKILL.md-compatible coding agents and supports shared state for multi-agent work.

Frontend Slides helps non-designers build web presentations by generating visual style previews first and letting the user choose a direction before the full deck is produced. It can also extract text, images, and notes from a PowerPoint file and rebuild them as a 16:9 HTML presentation with inline CSS and JavaScript. Optional scripts deploy the deck or export a static PDF; PowerPoint conversion requires Python and python-pptx.

This collection packages hundreds of agent skills, plugins, personas, commands, and scripts across engineering, DevOps, security, marketing, compliance, research, leadership, and everyday productivity. The skills work natively or through conversion with Claude Code, Codex, Gemini CLI, Cursor, and other coding agents. It also documents orchestration patterns for chaining skills or handing work between specialist personas.

opcode provides a desktop command center for Claude Code, combining project and session management with custom or background agents, usage tracking, MCP server controls, checkpoints, and CLAUDE.md editing. On first launch it detects the existing ~/.claude directory, so users can keep their native Claude Code configuration while working visually. The Claude Code CLI must still be installed and available on the system.

Set up OpenClaw, Hermes, or LangChain Deep Agents inside NVIDIA OpenShell with a reference stack that combines guided onboarding, routed inference, network policy, and sandbox lifecycle management behind one CLI. The project focuses on safer long-running agent operation and controlled egress. It is currently described as alpha, with community support provided on a best-effort basis.

Give a capable coding agent an article, argument, or product brief and this skill guides it through an editorial-magazine or Swiss-grid presentation workflow. It can produce a horizontal single-file HTML deck, supporting imagery, and matching social covers, with a low-power static mode for presentation. The format suits talks and opinion-led storytelling more than dense tables or collaborative editing, and its primary deliverable is HTML rather than native PPTX.

Rasa Open Source combines natural-language understanding with dialogue management for assistants that handle multi-turn conversations and connect to channels such as Slack, Telegram, and Twilio. The classic framework is now in maintenance mode, while newer Rasa agent development is moving toward Hello Rasa and the CALM engine.

Google ADK is a code-first Python toolkit for building, evaluating, and deploying AI agents. Its Agent and Workflow APIs support graph-based routing, nested workflows, and structured agent-to-agent tasks, while the CLI can run an agent locally or open a web UI for an agent directory. Version 2.0 changes the agent API, event model, and session schema, so existing applications need an upgrade check.

PentAGI lets autonomous agents plan and execute authorized penetration-testing work inside an isolated Docker environment. It includes more than twenty security tools such as nmap, Metasploit, and sqlmap, and can retain useful findings for later tasks. Cloud and local models are supported; smaller open models benefit from execution monitoring and task planning at the cost of higher token usage.

SWE-agent gives a selected language model tools to inspect a repository, edit code, and attempt a fix from a GitHub issue. It also supports custom coding tasks, SWE-bench experiments, and offensive-security challenges. The maintainers recommend the simpler mini-SWE-agent for most new users, while the EnIGMA mode currently directs users to SWE-agent 0.7.

Orca is an agent development environment for coordinating multiple terminal-based coding agents across separate worktrees. It brings repositories, terminals, previews, agent state, account switching, and usage tracking into one workspace. A mobile companion can notify you when work finishes, show progress, and send follow-up instructions while the desktop sessions continue running.

elizaOS combines a personal-assistant app with an agent runtime that can operate on desktop, mobile, web, a Linux desktop, or as an Android system assistant. It supports chat, voice, messaging, browser automation, on-device models, and persistent plugin-based apps while keeping the agent, data, and models local. Its optional cloud service adds hosted inference, cross-device sync, deployment, and metering rather than being required for local use.

Agent Zero runs an agent inside a Dockerized Linux desktop, where it can use terminals, browsers, desktop software, repositories, and editable documents, spreadsheets, and presentations. Projects isolate files, memories, secrets, and model settings, while multi-agent delegation, MCP, a host-machine bridge, and more than 100 community plugins extend the workspace. When connecting it to real host repositories, carefully limit the paths and permissions it can reach.

Pydantic AI is a Python framework for building agents with typed instructions, dependencies, tool calls, and structured outputs. It supports many model providers, MCP integrations, streamed results, tool-call approval, durable execution, and graph-based workflows. This makes it a fit for developers who want agent behavior checked by Python types and evaluated over time.