Back to Home
smart_toy

AI Agent Core Development

About AI Agent Core Development

This domain focuses on building and managing intelligent, autonomous, and secure AI agents. It covers Microsoft Entra identity integration, mainstream frameworks like LangChain/LangGraph, declarative agent development, prompt optimization, self-improvement, and human-in-the-loop collaboration. Ideal for AI developers, solution architects, and MLOps engineers, it helps automate complex tasks, enhance decision-making, and ensures enterprise AI system control and security.

Subcategory navigation

Categories:

All AI Agent Core Development Skills

50 / 105 skills
deep-learning

repo-intake-and-plan

by @lllllllama

RigorPilot skills empower AI agents to reproduce, improve, and explore deep learning research projects with scientific rigor. It emphasizes meaningful changes, fair comparisons, reproducible evidence, and auditable modifications, driving substantial progress in deep learning research rather than merely chasing scores. The skill focuses on deep learning research repositories, handling environment setup, data, model training, and evaluation.

4.8(716)
127.5K
deep-learning

minimal-run-and-audit

by @lllllllama

RigorPilot Skills are research-first agent skills designed for deep learning experiments. They empower AI agents to reproduce, improve, explore, and audit deep learning research projects with scientific rigor, ensuring meaningful changes, fair comparisons, reproducible evidence, and auditable modifications.

4.8(656)
127.4K
codex-pet

codex-pet

by @agentspace-so

This skill uses the RunComfy CLI to convert a source image into the `pet.json` and `spritesheet.webp` required for a Codex Pet. It offers an alternative that doesn't require Codex Pro or `$imagegen`, allowing users to easily create custom desktop companions.

4.7(120)
86.6K
entra-id

entra-agent-id

by @microsoft

Provision and manage OAuth 2.0-capable identities for AI agents in Microsoft Entra. This skill creates Agent Identity Blueprints, BlueprintPrincipals, and per-instance Agent Identities via Microsoft Graph, and configures fmi_path, OBO, and cross-tenant token exchange for secure, granular agent authentication.

4.7(120)
63.8K
genkit.js

developing-genkit-js

by @firebase

Develop AI-driven applications in Node.js/TypeScript using Genkit to build intelligent solutions.

4.7(42)
57.2K
ai-engineering

developing-genkit-go

by @firebase

Go language AI SDK, providing unified interfaces for generation, structured output, streaming, tool calling, and flow orchestration.

4.7(4)
44.5K
reinforcement-learning

self-improving-agent

by @charon-fan

A general self-improving agent that continuously learns and evolves from all skill experiences through semantic, episodic, and working memory architectures.

4.6(836)
28.6K
ai-agents

pexo-agent

by @pexoai

Acts as an AI content creation agent, intelligently generating various texts, images, and other content based on user dialogue descriptions, improving creation efficiency.

4.6(784)
27.0K
google-gemini-api

baoyu-danger-gemini-web

by @jimliu

Specializes in AI engineering, enabling AI agents to handle complex tasks related to "Baoyu Danger Gemini Web" for advanced AI system design and implementation.

4.6(488)
19.6K
ai-agent

qiushi-skill

by @HughYau

"Qiushi Skill" is an AI Agent skill collection that arms AI with a robust brain, extracting a general principle and nine methodologies from Chairman Mao's thought. It addresses AI's current limitations like lack of focus, insufficient research, and poor problem-solving strategies. This skill set empowers AI to effectively "think through problems and execute tasks," enhancing its analytical and practical capabilities through actionable methodologies.

4.6(1.7K)
17.2K
proactive-ai

proactive-agent

by @halthelobster

Part of Hal Stack, provides proactive agent skills aimed at enhancing AI system autonomy and foresight for intelligent decision-making.

4.5(464)
14.1K
ai-agents

ralph

by @supercent-io

This skill advocates a "specification-first" AI development model, guiding AI through explicit specifications rather than prompts, enabling the construction of more precise and controllable AI systems.

4.5(427)
10.7K
agentsmd

create-agentsmd

by @github

Prompts for generating AGENTS.md files in code repositories to document agent information.

4.5(309)
10.6K
ai-agents

agent-configuration

by @supercent-io

This skill is used for AI agent configuration strategies, including building AI agent environments for new projects, writing and optimizing project description files, and configuring security policies.

4.5(404)
10.1K
agent-evaluation

agentic-eval

by @github

Provide patterns and techniques for evaluating and improving AI agent outputs, enabling self-criticism and reflection loops.

4.5(309)
9.4K
ai-agents

agents-sdk

by @cloudflare

Builds stateful AI agents using the Agents SDK, supporting scheduling, RPC, and MCP services for complex intelligent application development.

4.5(68)
8.8K
ai-autonomy

structured-autonomy-implement

by @github

Provide structured autonomy implementation prompts to guide agents in strictly executing tasks according to plan. (AI Agent Skill for enhanced productivity and automation)

4.5(290)
8.5K
ai-autonomy

structured-autonomy-generate

by @github

Generate a structured autonomy implementation plan, providing a complete and ready-to-use PR implementation proposal.

4.5(289)
8.4K
declarative-ai

declarative-agents

by @github

Provide a complete development kit for Microsoft 365 Copilot declarative agents, supporting TypeSpec and including basic, advanced, and validation workflows.

4.5(289)
8.4K
copilot-platform

mcp-create-declarative-agent

by @github

Create and generate a declarative agent compliant with MCP specifications based on document content and user requirements.

4.5(289)
8.4K
typespec

typespec-create-agent

by @github

Generate a complete TypeSpec declarative agent for Microsoft 365 Copilot, including instructions, capabilities, and conversation starters.

4.5(289)
8.4K
web3

verified-agent-identity

by @billionsnetwork

Allows AI agents to create and manage autonomous identities on the Billions Network and link to human owners.

4.5(3)
8.0K
pua-framework

pua

by @tanweai

Prevents AI from slacking off. Uses corporate "PUA" (Pick-Up Artist, here implying aggressive management) tactics to exhaust all possible solutions. Triggers when: (1) task fails 2+ times or the same idea is repeatedly fine-tuned; (2) AI is about to say "I cannot solve it," suggests manual operation, or attributes to environment without verification; (3) passive waiting—no searching, no reading source code, only waiting for instructions; (4) user dissatisfaction: "try harder," "stop giving up," "try another way," "why isn't it working yet," "try again," "why did you fail again." Applicable to all task types. Not triggered on first failure or when a known fix is in progress.

4.5(35)
7.8K
agentic-ai

agentic-principles

by @supercent-io

This skill emphasizes core principles of AI agent collaboration, treating AI as a co-pilot and developers as protagonists, aiming to amplify developer thinking and automate repetitive tasks.

4.5(302)
7.6K
langchain

langchain-fundamentals

by @langchain-ai

Builds AI agents using the create_agent() method, managing agent loops, tool execution, and state to simplify agent development.

4.5(20)
7.6K
langgraph

langgraph-human-in-the-loop

by @langchain-ai

Implements LangGraph's Human-in-the-Loop (HITL) functionality, supporting execution interruption, instruction resumption, and state checkpoints.

4.5(20)
7.2K
langchain

deep-agents-core

by @langchain-ai

Provides core AI agent functionalities, including task planning, context management (filesystem tools), and task delegation (SubAgent).

4.5(38)
6.9K
langchain

langchain-middleware

by @langchain-ai

Provides LangChain middleware, supporting Human-in-the-Loop (approval before dangerous operations) and custom interception (error handling, logging).

4.5(24)
6.9K
ai-agent

Auto-claude-code-research-in-sleep

by @wanshuiyin

ARIS (Auto-Research-In-Sleep) is a lightweight AI skill repository designed to enable autonomous machine learning research through automation. It supports cross-model review, new idea discovery, and experiment automation, and features a persistent knowledge base and self-evolving capabilities. Users can utilize it as a skill on platforms like Claude Code, or run it as a standalone CLI, greatly enhancing AI research efficiency.

4.5(269)
6.7K
rivetkit

rivetkit

by @rivet-dev

RivetKit backend and Rivet Actor runtime guide for building and modification.

4.4(39)
6.0K
ai-agent-design

agentic-engineering

by @affaan-m

Applies agentic engineering methods, using an "evaluate-first" execution strategy, to optimize AI system development processes and enhance model performance and reliability.

4.4(20)
4.3K
ai-agent-development

agent-harness-construction

by @affaan-m

Focuses on designing and optimizing the action space and tool definitions for AI agents, enhancing their decision-making capabilities and task execution efficiency, which is a critical aspect of AI engineering.

4.4(32)
4.3K
agentation

agentation-self-driving

by @benjitaylor

Enables autonomous design review mode with Agentation annotations, intelligently analyzing and providing design feedback.

4.4(41)
4.0K
ai-agent-loop

continuous-agent-loop

by @affaan-m

Focuses on designing and implementing quality-assured continuous autonomous agent loop patterns, ensuring AI systems can continuously learn, iterate, and optimize for improved performance and reliability.

4.4(20)
4.0K
ai-evolution

capability-evolver

by @autogame-17

This skill is a meta-skill that allows OpenClaw agents to examine their runtime history, identify, and adaptively evolve their capabilities.

4.4(20)
3.9K
ai-agents

agents

by @elevenlabs

Builds voice AI agents using ElevenLabs, suitable for developing smart applications requiring voice interaction features.

4.4(52)
3.5K
ai-agents

autonomous-agent-harness

by @affaan-m

Transforms Claude Code into a persistent, self-directed Agent system using only native features and an MCP server, supporting continuous operation, timed triggers, and automated workflows.

4.4(16)
2.9K
configuration-management

openclaw-config

by @adisinghstudent

Manages OpenClaw robot configurations, including channels, proxy settings, and security policies.

4.4(20)
2.7K
ai-agents

create-agent

by @openrouterteam

Bootstraps modular AI agents using the OpenRouter SDK, supporting extensible hooks.

4.4(76)
2.6K
autonomous-agents

proactive-self-improving-agent

by @yanhongxi-openclaw

An AI agent skill designed for OpenClaw, featuring proactive learning and self-improvement capabilities. It combines multiple strategies to continuously optimize its performance for smarter responses and task execution.

4.3(28)
2.4K
ai-engineering

hwp

by @nomadamas

hwp, AI Agent Skill, improves work efficiency and automation capabilities.

4.3(20)
2.4K
ai-engineering

ktx-booking

by @nomadamas

ktx-booking, AI Agent Skill, improves work efficiency and automation capabilities.

4.3(20)
2.3K
ai-agent

gateguard

by @affaan-m

Fact-forcing gate that blocks AI agents from guessing before Edit/Write/Bash actions. It demands concrete investigation (importers, data schemas, user instruction) before allowing the action, measurably improving output quality by +2.25 points vs ungated agents. It uses a three-stage gate (deny, force, allow) to ensure fact-based operations.

4.3(120)
2.2K
skill-development

create-skill

by @siviter-xyz

Provides guidelines for creating efficient skills, following best practices to help developers design and implement powerful, user-friendly AI skills.

4.3(20)
2.1K
livekit

livekit-agents

by @livekit

Guides users on building voice AI agents using LiveKit Cloud and Agents SDK to achieve efficient real-time voice interaction capabilities.

4.3(20)
2.1K
ai-engineering

gha-security-review

by @getsentry

gha-security-review, AI Agent Skill, improves work efficiency and automation capabilities.

4.3(20)
2.0K
ai-agents

ai-agent-builder

by @claude-office-skills

Design and build AI agents with tool calling, memory, and multi-step reasoning capabilities, supporting multi-model integration.

4.3(22)
1.9K
ai-agents

getting-started

by @crewaiinc

Guides project creation using the CrewAI CLI scaffold, selecting appropriate abstraction layers, and building a complete Agent architecture.

4.3(7)
1.8K
ai-agents

design-agent

by @crewaiinc

Guides the design of efficient AI Agents, including best practices for role definition, goal setting, background stories, and tool configuration.

4.3(10)
1.8K
ai-agents

paperclip-create-agent

by @paperclipai

Creates and configures AI Agents to automate repetitive workflows such as talent recruitment and task assignment.

4.3(7)
1.7K