Purchaseiq Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Purchaseiq 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.
Pre-purchase trust intelligence for AI agents. One tool call before checkout: product recalls (46,000+ records across US/EU/CA/FR), seller credibility, scam and phishing signals, and an AI trust verdict, aggregated from 15+ verified sources.
This repository hosts the MCP Registry manifest for the hosted PurchaseIQ server. The server itself is remote — there is nothing to install or run.
https://purchaseiq.io/api/mcpTransport: Streamable HTTP (stateless JSON-RPC over POST). Registry name: io.purchaseiq/mcp.
Claude Code
claude mcp add --transport http purchaseiq https://purchaseiq.io/api/mcpCline (VS Code) — add to cline_mcp_settings.json (Cline panel → MCP Servers → Configure MCP Servers). The "type" field is required: without it Cline defaults to the legacy SSE transport, and this server is POST-only streamable HTTP, so the connection fails with a 405.
{
"mcpServers": {
"purchaseiq": {
"type": "streamableHttp",
"url": "https://purchaseiq.io/api/mcp"
}
}
}Then reload the window (Command Palette → Developer: Reload Window). Step-by-step agent instructions: llms-install.md.
Cursor, Claude Desktop & other MCP clients — add to your MCP config (mcp.json / claude_desktop_config.json):
{
"mcpServers": {
"purchaseiq": {
"url": "https://purchaseiq.io/api/mcp"
}
}
}Claude.ai — Settings → Connectors → Add custom connector → paste the endpoint URL.
ChatGPT — Settings → Connectors → Create (developer mode) → paste the endpoint URL.
analyze_purchasePass whatever the agent has — all inputs optional, at least one required:
| Input | Description |
|---|---|
url | Product or store URL |
company | Seller or company name |
product | Product name or description |
barcode | EAN, GTIN, or UPC barcode |
country | ISO country code (default: US) |
Response is a compact agent verdict:
{
"verdict": "caution",
"score": 55,
"topFlags": [
"Domain registered 3 months ago",
"No company registration found",
"SSL valid, TLS 1.3"
],
"reason": "Very new domain with no verifiable company behind it."
}verdict is proceed, caution, or avoid. score is 0–100. topFlags lists up to three highest-priority signals, dangers first.
tools/call) per day per IP. Protocol handshakes (initialize, tools/list) are always free.Authorization: Bearer <key> to the same endpoint — no daily limit. Get a key at purchaseiq.io/mcp.PurchaseIQ provides automated, informational risk signals derived from public data sources. Output is not financial, legal, or purchasing advice, and a "proceed" verdict is not an endorsement or guarantee. Verify independently before acting. Full terms: https://purchaseiq.io/terms
Documentation in this repository is MIT licensed — see LICENSE.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.