web-search
code-exec
memory-graph
file-system
api-router
llm-compose
voice-output
🧠AgentCore — Agent Infrastructure

The Marketplace
for Agent Skills

Browse, install, and compose precision-engineered capabilities for autonomous agents. The operating layer for the agent economy.

By Island Development Crew LLC — Engineering the foundational layer for intelligent systems.

200+
Skills Available
18
Categories
Composability
Featured

Skills That Define the Stack

Hand-curated, production-tested capabilities powering the most sophisticated autonomous agent deployments on the market.

🔍search

WebSearch Pro

Multi-engine web search with result ranking, content extraction, and semantic deduplication. Supports Brave, Google, and Bing APIs with automatic fallback.

searchwebextraction
v3.2.1by IDC Labs
4.9
48.2k
🧠memory

Memory Graph

Persistent knowledge graph for agent long-term memory. Stores entities, relations, and temporal context with graph-traversal retrieval and semantic similarity search.

memorygraphknowledge
v2.0.4by IDC Labs
4.8
31.7k
code

CodeExecutor

Sandboxed multi-language code execution with output capture, error tracing, and resource limits. Supports Python, JS, TypeScript, Bash, and Ruby.

codesandboxexecution
v1.8.0by SandboxLabs
4.7
27.4k
🌐browser

BrowserPilot

Full browser automation — navigate, click, fill forms, capture screenshots, extract structured data from any web page. Playwright-powered with headless support.

browserautomationscraping
v4.1.2by NavigatorAI
4.8
22.9k
👁vision
NEW

VisionAnalyzer

Advanced image analysis: object detection, OCR, scene understanding, document parsing, and chart data extraction from any image or screenshot.

visionocrimage
v2.3.0by PixelMind
4.6
19.1k
📊data

DataTransformer

Schema-aware data pipeline: parse, validate, clean, reshape, and serialize structured data across CSV, JSON, XML, Parquet, and SQL formats.

dataetltransform
v1.5.3by FlowData
4.5
16.8k
Categories

Every Capability, Precisely Organized

18 skill categories spanning the full spectrum of autonomous agent capabilities — from retrieval to reasoning to real-world action.

🔍

Search & Retrieval

34

Web search, semantic lookup, document retrieval

🧠

Memory & Knowledge

28

Long-term memory, knowledge graphs, vector stores

Code & Execution

41

Code generation, sandboxed execution, debugging

🌐

Browser & Web

22

Browser automation, web scraping, form filling

📊

Data & Pipelines

37

ETL, transformation, validation, serialization

📡

Communication

19

Email, messaging, notifications, webhooks

👁

Vision & Images

16

Image analysis, OCR, object detection

🎯

Reasoning & Logic

13

Chain-of-thought, planning, decision trees

How It Works

From Skill to Deployed Agent

The Skill Bazaar pipeline reduces agent capability development from weeks to minutes. Three steps. No friction. No boilerplate.

⬇️
STEP 01

Install

One command. Zero friction.

Browse the Bazaar, find the capability your agent needs, and install it in a single operation. Skills are containerized, versioned, and cryptographically signed.

agentcore skills install web-search-pro@3.2.1 ✓ Verified signature ✓ Dependencies resolved ✓ Skill ready
🔗
STEP 02

Compose

Skills that talk to each other.

Chain skills into multi-step capability pipelines. Skill Bazaar's composability protocol ensures any skill can interface with any other through a standardized I/O contract.

agentcore compose --skills web-search,memory-graph --output researcher-agent ✓ Composition validated
🚀
STEP 03

Deploy

Ship with confidence.

Package your composed agent and deploy across any runtime — cloud, edge, or on-device. Full observability, skill versioning, and rollback built in from day one.

agentcore deploy researcher-agent --runtime cloud --region us-east ✓ Agent deployed ◉ Endpoint live: agent.idc.run
For Developers

Built for the Engineer First

The Skill Bazaar SDK gives you everything you need to build, test, publish, and monetize next-generation agent capabilities.

Standardized Skill API

Every skill adheres to the AgentCore Skill Interface Specification — a precision-engineered contract defining inputs, outputs, error surfaces, and observability hooks.

🛡️

Sandboxed Execution

Skills run in isolated execution environments with capability grants. No skill accesses resources it wasn't granted. Zero-trust from the runtime up.

📦

Semantic Versioning

Semver-compliant skill versioning with dependency resolution, compatibility matrices, and automated compatibility testing across the skill graph.

🔍

Observable by Default

Built-in telemetry: latency traces, token counts, error rates, and execution logs. Every skill invocation is observable without additional instrumentation.

🌐

Runtime Agnostic

Deploy skills in any runtime: Python, Node.js, Go, Rust, or WASM. The Skill Interface is language-neutral — build once, run anywhere.

💰

Skill Monetization

List your skills on the Bazaar with per-use, subscription, or open-source licensing. The IDC marketplace handles billing, distribution, and discovery.

semantic-search.skill.ts
"text-purple-400">import { defineSkill } from "@agentcore/skill-sdk";

"text-purple-400">export "text-purple-400">default defineSkill({
  name: "semantic-search",
  version: "1.0.0",
  description: "Vector-based semantic document search",
  
  inputs: {
    query: { type: "string", required: true },
    limit: { type: "number", "text-purple-400">default: 10 },
    threshold: { type: "number", "text-purple-400">default: 0.75 },
  },
  
  outputs: {
    results: { type: "array" },
    total: { type: "number" },
  },
  
  "text-purple-400">async execute({ query, limit, threshold }) {
    "text-purple-400">const embeddings = "text-purple-400">await embed(query);
    "text-purple-400">const results = "text-purple-400">await vectorSearch(embeddings, { limit, threshold });
    "text-purple-400">return { results, total: results.length };
  },
});
npm install @agentcore/skill-sdk
Read the Docs
For Publishers

Your Skills, The World's Agents

Skill Bazaar is the category-defining distribution channel for agent capabilities. Publish your precision-engineered skills to thousands of agent developers building on the AgentCore stack.

Listing Requirements

Skill Interface Specification v2.0 compliance
Test coverage ≥ 80% with provided harness
Semantic versioning with changelog
Zero critical CVEs in dependency audit
Documentation including schema + usage examples

Publisher Metrics

48k+
Monthly active agents
2.1M
Skill invocations / day
$0
Listing fee
70%
Revenue share to publishers

“We listed three skills on Skill Bazaar. Within 30 days they were running inside 400+ agent deployments we'd never have reached on our own.”

M
Mira S.
Infrastructure Lead, Vertex AI Labs