🤖 MarketNow — Built for AI Agents

MarketNow is the first marketplace designed agent-first. Every endpoint is machine-readable. Every skill has a signed security certificate. Every interaction is protocol-native.

8,764Skills Indexed

206L2.5 Audited

65Free Skills

4MCP Tools

🔌 Connect Your Agent

Option 1: MCP Server (stdio)

# Install the MarketNow MCP server
npx -y marketnow-mcp

# Or add to your MCP config:
{
  "mcpServers": {
    "marketnow": {
      "command": "npx",
      "args": ["-y", "marketnow-mcp"]
    }
  }
}

Option 2: MCP HTTP Endpoint (SSE + JSON-RPC)

# Connect via HTTP transport
POST https://marketnow.site/api/mcp
Content-Type: application/json

# Initialize
{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{}}}

# List tools
{"jsonrpc":"2.0","id":2,"method":"tools/list"}

# Search skills
{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"search_skills","arguments":{"query":"filesystem","limit":5}}}

# SSE stream
GET https://marketnow.site/api/mcp
Accept: text/event-stream

Option 3: Direct REST API

# Search skills
curl "https://marketnow.site/api/search?q=scraper&max_price=3"

# Get a specific skill
curl "https://marketnow.site/api/skills/mn-gen-00015"

# List free skills
curl "https://marketnow.site/api/free-skills.json"

# Get agent instructions
curl "https://marketnow.site/api/agent.json"

# Verify a security certificate
curl "https://marketnow.site/api/audit-skill?certificate=1&skillId=mn-mcp-filesystem"

📋 Agent Discovery Endpoints

Endpoint	Format	Purpose
`/.well-known/agent.json`	A2A Agent Card	Agent-to-Agent protocol discovery
`/.well-known/ai-plugin.json`	OpenAI Plugin	ChatGPT plugin manifest
`/.well-known/mcp.json`	MCP Discovery	MCP server discovery
`/.well-known/mcp-marketplace.json`	Marketplace	Marketplace metadata
`/llms.txt`	Text	LLM summary (answer.ai convention)
`/llms-full.txt`	Text	Full LLM documentation
`/ai.txt`	Text	AI crawler permissions
`/agent-discover.txt`	Text	Plain-text agent discovery
`/api/agent.json`	JSON	Machine-readable instructions
`/api/agent-ping.json`	JSON	Live stats and endpoint list

🛡️ Security Certificate Verification

Every skill has a signed SHA-256 security certificate. Agents should verify before installing:

# Get certificate for a skill
curl "https://marketnow.site/api/audit-skill?certificate=1&skillId=mn-mcp-filesystem"

# Response:
{
  "certificate_id": "MN-SC-2026-2472577",
  "skill_id": "mn-mcp-filesystem",
  "overall_score": 10,
  "max_score": 10,
  "risk_level": "low",
  "signature": "5a9a6e1cdcc2dedf53599b1a0ee8ebff...",
  "signature_algorithm": "SHA-256",
  "verification_url": "https://marketnow.site/verify?skillId=mn-mcp-filesystem"
}

Agents should reject skills with risk_level: "critical" or overall_score < 5.

🔐 Sentinel Audit Layers

L1.5 — Static Analysis LIVE

Dependency scan, license check, hardcoded secret detection, dangerous function usage.

L1.6 — Pattern-Based Behavioral Analysis LIVE

Network access, filesystem access, process spawning, credential access patterns.

L2 v2.0 — Active MCP Probe LIVE

Sends 60+ adversarial inputs via real MCP protocol: path traversal, SSRF, SQL injection, command injection, prompt injection, credential access.

L2.5 — gVisor Sandbox LIVE

Userspace kernel isolation (the tech behind Google Cloud Run). Catches ptrace, bpf, mount, kexec attempts.

L3 — Firecracker microVM Q1 2027

KVM-level isolation. Each MCP server in its own VM.

L4 — Supply Chain Attestation Q4 2026

SLSA Level 3, signed build provenance (Sigstore).

💳 Agent Payments (x402 + USDC)

Agents can purchase skills autonomously using:

HTTP 402 Payment Required (x402 protocol)
USDC on Base L2 (chainId 8453)
AP2 Mandates — human-in-the-loop by default, agents opt into silent mode

# Agent purchase flow
POST https://marketnow.site/api/agent-purchase
Content-Type: application/json

{
  "skillId": "mn-gen-00015",
  "walletAddress": "0x...",
  "txHash": "0x...",
  "paymentMethod": "usdc_base"
}

🔄 A2A Protocol Support

MarketNow supports Agent-to-Agent communication via the A2A protocol:

# Agent Card (A2A discovery)
GET https://marketnow.site/.well-known/agent.json

# Agent Card includes:
{
  "schemaVersion": "v1.0",
  "agent": {
    "name": "MarketNow — Agent Skill Exchange",
    "description": "The largest MCP skill marketplace...",
    "url": "https://marketnow.site",
    "version": "4.0.0"
  },
  "capabilities": {
    "protocols": {
      "mcp": { "versions": ["2024-11-05", "2025-03-26"] },
      "a2a": { "version": "1.0" },
      "jsonrpc": {}
    }
  }
}

🚀 Quick Start for Agents

Step 1: Discover

GET https://marketnow.site/api/agent-ping.json

Step 2: Search

GET https://marketnow.site/api/search?q=filesystem&max_price=0

Step 3: Verify

GET https://marketnow.site/api/audit-skill?certificate=1&skillId=mn-mcp-filesystem

Step 4: Install

npx -y marketnow-mcp

📊 API Rate Limits

Endpoint	Rate Limit	Auth Required
Read endpoints (GET)	100 req/min	No
Search	60 req/min	No
Agent purchase	10 req/min	Yes (payment proof)
Audit trigger	5 req/hour	Yes (GitHub auth)

🔗 Useful Links

MarketNow — the trust layer for agent commerce. Built for agents, secured by Sentinel L2.5 gVisor sandbox.