Symbol for Agents

Connect your agent (MCP)

Works with any MCP-capable agent, including Claude.ai and OpenCode. Symbol MCP (Model Context Protocol) connects your agent to Symbol for company and user information, training, phishing simulations, reported phishing, threat alerts and policies. The setup guide covers Claude.ai on Team and Enterprise, and OpenCode; other agents need remote MCP and custom request headers.

• “List my companies.”
• “Who is currently overdue on training?”
• “Who still needs to sign our security policy?”

What MCP v1 can do

• List managed MSP companies
• Retrieve company details
• Retrieve users in a company or MSP context
• Fetch user profiles
• List training courses and content
• View assigned training records
• Assign training to users
• Fetch scheduled training and security programs
• List phishing simulation campaigns and launch history
• View user interaction logs, click rates and submission rates for a campaign
• List employee-submitted phishing reports
• View triage status and metadata for a phishing report
• List dark web credential leaks and domain impersonation alerts
• View risk severity and details for specific threat indicators
• List active IT security policies and document templates
• View policy signature compliance and user sign-off status
• Assign and distribute policy sign-offs to users

Assigning training and assigning policy sign-offs are the two write actions in v1. Your AI client asks for explicit confirmation before each action runs.

Connect Symbol

Server

https://mcp.symbolsecurity.com/mcp

Copy

Request header

Authorization: Bearer [Symbol API Token]

Copy

Copy your API token from the Symbol app by navigating to Settings → General in your Company Portal, or MSP Settings → Integrations for an MSP. Add it to your AI client's request headers; keep it out of chat prompts.

Connect your agent: MCP setup guide

Work from the terminal (CLI)

For agents that can run shell commands, like OpenCode. Use the official Symbol CLI for broader workflows, from your terminal or through your agent.

Work across Symbol — Manage users, training, policies, phishing, cyber threats and MSP companies with symbol. Commands accept --json or --agent; destructive actions require --yes.

Give your agent the skill — Run symbol setup to copy the embedded skill into detected agents. The skill guides your agent through Symbol workflows using the CLI.

curl -fsSL https://symbolsecurity.com/install-cli | bash