Shape agent capabilities into your workflow.

Claw Code Agent reimplements a Claude Code-style coding workflow in Python for OpenAI-compatible model servers. It provides a CLI, multi-turn chat, a local web interface, file and shell tools, sessions, permissions, plugins, and nested delegation while using only the Python standard library. Development is ongoing, and full MCP, hook, background-team, and platform parity is not yet complete.

Give coding agents versioned standards for languages, frameworks, and software delivery without replacing the tools your team already uses. The CLI initializes, syncs, and validates native assets for agents such as Cursor, Claude Code, Copilot, and Gemini, while an optional MCP server can load matched guidance at runtime and preserve project-specific overrides.

Power Design is a Claude Code skill that turns a short brief into either an HTML presentation or a responsive website. It first asks which format and brand to use, then can extract brand characteristics from a URL through Firecrawl or select from more than 72 bundled brand systems. Separate sets of 20 rules guide slide and web output. The resulting HTML opens in a browser and can be refined through conversation, making it useful when users want more deliberate structure and styling than a generic AI template.

Firecrawl CLI runs search, scraping, interactive browsing, mapping, crawling, and agent jobs from the terminal. Its setup commands can also install Firecrawl skills, workflow skills, and MCP configuration for Claude Code, Codex, Cursor, Windsurf, and other supported harnesses; authentication is required for service access, and anonymous telemetry can be disabled.

Aiden is a Windows-native, local-first autonomous agent that can operate a browser and terminal, invoke tools, run workflows, and persist memory and job state on the user's machine. It supports multiple model providers, subagents, scheduled or external triggers, and long-running work with verification, retries, recovery, and resume behavior. The project is released under AGPL-3.0.

AgentSight observes AI agents at the system level, correlating process trees, shell commands, file operations, network destinations, prompts, responses, tool intent, and model token data. It works outside closed-source CLIs without an SDK, proxy, or vendor integration, and offers live views, SQLite session reports, and offline pprof profiles. It requires Linux eBPF support and elevated probe privileges, and its capture scope includes potentially sensitive prompts and responses.

g3 is a Rust terminal agent with file, shell, desktop-control, code-search, and TODO tools. It connects to Anthropic, Databricks, or local llama.cpp models with Metal acceleration on macOS, monitors context usage, thins large tool results, and retries recoverable errors. Local models support offline or privacy-focused work, but the project notes weaker results on complex tasks and marks computer control as experimental.

Amazon Sorftime MCP Skills combines Sorftime, Sif, Xiyou Insight, SellerSprite, and related data services into nine workflows for listing analysis, category selection, keyword research, review mining, traffic, ads, pricing, and listing optimization. Reports can be produced in Markdown, CSV, Excel, or HTML formats. Running the workflows depends on access to the relevant MCP services, accounts, and data permissions.

Swift Concurrency Pro is an agent skill for Claude Code, Codex, Cursor, Gemini, and other coding assistants. It guides writing and review across async/await, actors, Sendable, task groups, structured concurrency, newer language features, and migration choices, with rules aimed at mistakes language models commonly make. After installation, users can invoke the skill with a dedicated command or natural-language request and narrow the review to a specific concern. It supplements an existing coding agent rather than running as a separate Swift development tool.

Catnip creates containerized workspaces with managed Git worktrees so multiple Claude Code sessions can run in parallel. Its web and native mobile interfaces, CLI, and SSH access expose full terminals, while port detection and a built-in proxy let you preview services from the workspace. It can run in Codespaces, devcontainers, or locally, but the current workflow is optimized for Claude Code and broader agent support is still planned.

Help an AI coding agent design tables, choose keys and data types, tune queries and joins, plan ingestion, or troubleshoot ClickHouse workloads using official guidance. The repository also covers ClickHouse Cloud and chdb for in-process analytics, and can be installed through the skills CLI or clickhousectl across multiple supported agents.

MCP-Zero is the open-source code accompanying a paper on active tool discovery and retrieval for autonomous LLM agents. It includes instructions for an MCP-tools dataset covering 308 servers and 2,797 tools, plus code under MCP-tools/build_data for constructing a custom server dataset. The data file must be obtained separately and placed at the path expected by the evaluation code.

rnskill collects project-level skills for Chinese prose editing, motion-first video direction, original reference-motion studies, dark SaaS product videos, typed text openers, and evidence-based replica quality control. It works with Codex, Claude Code, and other SKILL.md-compatible agents, and users can install the whole collection or select individual skills. The repository uses a CC BY-NC 4.0 license, so commercial use requires careful review of the license terms.

Produck Skills is an open-source bundle for coding agents that turns vague user requests into scoped, testable PRDs and carries real feedback into implementation work. The collection also includes workflows for producing animated product demos and improving website visual judgment through screenshot-based review. It can be installed as a bundle across Codex, Claude Code, Cursor, and other compatible agents. The feedback workflow needs a Produck MCP token to load ticket context, while the remaining skills can be used without it.

skills-npm lets packages ship matching Agent Skills alongside their code, then links those skills into detected coding-agent directories. A one-time setup adds a prepare hook, an ignore rule, and the first synchronization so future dependency installs refresh the links automatically. Configuration can limit packages, skills, agents, or monorepo traversal.

Agent Deck provides one terminal view for coding-agent sessions across many projects, making it easy to see which are running, waiting for input, or finished. Groups, search, session forking, and Git worktrees help organize parallel work, while the conductor can surface sessions that need attention remotely.

PM-Skills provides 68 reusable product-management skills for hypotheses, research, PRDs, user stories, Foundation Sprints, Design Sprints, and delivery artifacts. It includes orchestrated workflows, templates, and more than 200 sample outputs, with native Claude Code installation and cross-agent use through the open skills CLI for Codex, Cursor, Copilot, and others.

visualise teaches an agent to produce self-contained HTML and SVG for clickable diagrams, live charts, comparison views, steppers, and other explainers. Clients that render its visualizer code fence can display the result inline in a sandboxed frame. In unsupported clients, the generated HTML remains usable as a standalone browser file.