Agent Central Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Agent Central Mcp (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
agentcentral is a hosted Amazon MCP server for Amazon sellers: an Amazon Seller Central MCP server and Amazon Ads MCP server for Claude, ChatGPT, Cursor, Claude Code, OpenClaw, and other HTTP-capable MCP clients.
It connects AI clients to Amazon Ads, Seller Central, inventory, orders, catalog, rankings, finance, and fulfillment data through the hosted Streamable HTTP endpoint at https://mcp.agentcentral.to/mcp.
This is not just an export layer. agentcentral is an operational MCP server with fast factual reads plus safe guarded writes: scoped API keys, read/write tool separation, previews, guardrails, and audited write results. The hosted endpoint currently exposes 141 domain-scoped tools and 145 production tools total. This public repo contains registry metadata and a thin stdio introspection stub; live tool execution happens only through the hosted endpoint. The local stub lists the hosted catalog plus one setup helper for discovery, so it is intentionally one tool larger than the hosted production catalog.
{
"mcpServers": {
"agentcentral": {
"url": "https://mcp.agentcentral.to/mcp",
"headers": {
"Authorization": "Bearer ac_live_<YOUR_API_KEY>"
}
}
}
}Claude custom connectors use a signed connector URL generated in the agentcentral dashboard.
Full setup guide: https://agentcentral.to/amazon-seller-central-mcp-claude
agentcentral returns factual seller data, source fields, deterministic metrics, classifications, and audited write results. It is not a recommendation engine.
This repo ships a minimal stdio MCP server so directories and clients can introspect the public tool catalog without an agentcentral account. It does not execute tool calls. Every call returns a pointer to the hosted endpoint and setup guide.
The stub is not published to npm. Clone this repository or use the Docker example below if a directory/client needs local stdio introspection; for real usage, configure the hosted remote MCP endpoint above.
docker build -t agentcentral-mcp .
docker run --rm -i agentcentral-mcpOr via Node:
npm install
npm run build
node dist/index.js~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.