agentaegis-security

AgentAegis security scans

AgentAegis is a hosted MCP server that runs real security scanners (nmap, Nuclei, sslyze, Semgrep, trufflehog, trivy) and queries real threat-intel feeds (NVD, AbuseIPDB, AlienVault OTX, abuse.ch, HIBP). Findings describe real infrastructure.

Setup

Add the server to your MCP client:

{
  "mcpServers": {
    "agentaegis": {
      "url": "https://agentaegis-mcp-production.up.railway.app/mcp",
      "headers": { "Authorization": "Bearer aegis_YOUR_API_KEY_HERE" }
    }
  }
}

Get a key at https://agentaegis.org/pay. Alternatively, omit the header and pay per call in USDC on Base via x402 — no signup, the server answers with an HTTP 402 challenge your x402-capable client settles automatically.

If the tools are not available, the server is not connected. Say so rather than guessing at answers this skill is meant to measure.

Cost — this skill spends real money

Tools cost $1 to $5 per call, charged when the call is made. A scan that returns nothing useful still costs.

• Before the first paid call in a conversation, tell the user which tool you plan to call and what it costs, and wait for their agreement.
• Never run a scan the user did not ask for. Never call paid tools to explore the catalog.
• Never retry a paid tool in a loop. On error, report it and stop.

Free, always: help, account_balance, agent_whoami, agent_history, agent_scan_get.

Call help for exact current pricing rather than quoting numbers from memory.

Always check for a prior result first

Every paid scan is stored. agent_history lists your prior scans and agent_scan_get returns any one of them in full, both free. When the user asks about a target you have looked at before, retrieve it instead of re-paying. Only re-scan when they want fresh data or the earlier scan is stale.

Choosing a tool

The user is asking	Use
Is this endpoint/domain safe to call, trust or pay?	vet_endpoint
Should I install this MCP server or plugin?	scan_mcp_plugin
Should I trust this agent skill?	scan_skill
What is CVE-YYYY-NNNNN?	cve_lookup
Is this IP or domain malicious?	threat_intel_lookup
Can our email be spoofed?	email_security_audit, or dns_security_check for records only
Is our TLS configured correctly?	ssl_tls_audit
Are there vulnerabilities in this code?	sast_scan
Are there hardcoded secrets?	secret_scan
Are our dependencies vulnerable?	dependency_audit
What ports are exposed?	vuln_scan_network
Is this web app vulnerable?	vuln_scan_web_app
Are we SOC 2 / ISO 27001 / HIPAA / PCI-DSS / NIST CSF ready?	compliance_framework_check
We have an incident	incident_triage