AI agents invoke caddy_reload_tool to trigger actions in Infra. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
The tool executes a Caddy reload operation, which is a command that triggers an external system (the Caddy reverse proxy) to reconfigure itself. While the confirmation requirement (confirm=true) mitigates risk, the tool still Execute-class because it initiates active service operations.
From the tool's definition 'Reload Caddy after a Caddyfile change' — reloading a reverse proxy service triggers external operations (Caddy process restart) whose effects depend on configuration arguments.
Attacks that exploit this kind of access
Reload Caddy after a Caddyfile change. Requires confirm=true. It is categorised as a Execute tool in the Infra MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Infra MCP server in PolicyLayer and add a rule for caddy_reload_tool: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Infra. Nothing to install.
caddy_reload_tool is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the caddy_reload_tool rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for caddy_reload_tool. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
caddy_reload_tool is provided by the Infra MCP server (odanree/infra-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →