Agentic AI refers to artificial intelligence capable of making decisions, planning, and adapting to new information in real time. AI agents learn and enhance their performance through feedback, utilizing advanced algorithms and sensory inputs to execute tasks and engage with their environments.

AI agent features three key characteristics:

  • Planning: an AI agent is capable of creating a step-by-step plan with discrete milestone goals from a prompt while learning from mistakes via a reward system to improve future outputs.
  • Memory: an AI agent combines the ability to use short-term memory to process chat-based prompts and follow-up prompts with longer-term data retention and recall (often RAG).
  • Tool use: an agent can query APIs to request additional information or execute an action based on an end user’s request.

Tools & Frameworks

ToolDescription
Browser-useMake websites accessible for AI agents
AgentariumOpen-source framework for creating and managing simulations populated with AI-powered agents
Jobs_Applier_AI_Agent_AIHawkJobs_Applier_AI_Agent_AIHawk aims to easy job hunt process by automating the job application process
SwarmsThe Enterprise-Grade Production-Ready Multi-Agent Orchestration Framework
MemaryThe Open Source Memory Layer For Autonomous Agents
Micro-agentAn AI agent that writes (actually useful) code for you
ACUA curated list of resources about AI agents for Computer Use, including research papers, projects, frameworks, and tools
LightpandaLightpanda: the headless browser designed for AI and automation
cogneeMemory for AI Agents in 6 lines of code
agent-lightningThe absolute trainer to light up AI agents
PaperclipOpen-source platform to orchestrate teams of AI agents with org charts, goal cascading, task ticketing, budget controls, and human approval gates
agency-agentsA complete AI agency at your fingertips — pre-built agent personalities (engineering, design, sales, marketing, product, QA) for Claude Code, Cursor, Aider, and Windsurf
gstackOpen-source software factory built on Claude Code with 23 specialised AI agents covering a full dev sprint cycle
ConductorMac app to run multiple coding agents (Claude Code, Codex) in parallel, each isolated in its own git worktree
SupersetDesktop IDE for running 10+ CLI coding agents simultaneously in isolated git worktrees, with diff viewer and review UI
vibe-kanbanKanban board and agent execution platform for managing tasks and reviewing AI coding agent output with branch isolation
IcarusShared memory system for AI agents that store structured markdown files, enabling agents to write, read, and search collaborative work history
KanbanWeb app for managing multiple CLI agents in parallel with kanban board interface, providing isolated worktrees and terminals for each agent
AgentMemoryPersistent, searchable memory database for AI coding agents that captures tool usage and injects relevant context automatically across sessions
GooseOpen-source AI agent that automates complex engineering tasks autonomously, supporting multiple LLMs and MCP server integration
Agentic AI Prompt ResearchDocuments reconstructed prompt patterns and architectural designs that power agentic AI coding assistants for building autonomous systems
OpenWorkOpen-source local-first platform for running agentic AI workflows with desktop app, CLI, and web server interfaces

Agent Skills

SkillDescriptionInstall
create-agentsmdGenerates standardized AGENTS.md files to help AI coding agents understand and work with your repository. Covers setup, workflow, testing, code style, and deployment. Compatible with 20+ AI coding tools.npx skills add https://github.com/github/awesome-copilot --skill create-agentsmd
agent-governanceDeclarative policies, intent classification, and audit trails for controlling AI agent tool access and behavior. Includes threat detection, rate limiting, trust scoring, and immutable audit logs.npx skills add https://github.com/github/awesome-copilot --skill agent-governance
agentic-evalIterative evaluation and refinement patterns for improving AI agent outputs through self-critique loops. Supports reflection, evaluator-optimizer architecture, and test-driven refinement.npx skills add https://github.com/github/awesome-copilot --skill agentic-eval
team-communication-protocolsStructured messaging protocols for coordinating agent teams through direct messages, broadcasts, plan approvals, and graceful shutdown. Includes teammate discovery and anti-pattern documentation.npx skills add https://github.com/wshobson/agents --skill team-communication-protocols
code-reviewerAutomated code review for local changes and remote PRs with structured analysis across correctness, maintainability, security, efficiency, and test coverage. Supports GitHub CLI checkout.npx skills add https://github.com/google-gemini/gemini-cli --skill code-reviewer