AI agents call phishfort_plan_change to retrieve information from Phishfort without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs no side effects; it reads/computes a plan without modifying any data. It is analogous to a dry-run or preview operation, placing it firmly in the Read category with low severity since misuse cannot directly cause harm.
From the tool's definition 'Plan a PhishFort write. Does not mutate.' — explicitly states no mutation occurs; it only plans/previews a change and returns approval fields for use with actual write tools.
Attacks that exploit this kind of access
Plan a PhishFort write. Does not mutate. Use returned approval fields on write tools. It is categorised as a Read tool in the Phishfort MCP Server, which means it retrieves data without modifying state.
Register the Phishfort MCP server in PolicyLayer and add a rule for phishfort_plan_change: 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 Phishfort. Nothing to install.
phishfort_plan_change is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the phishfort_plan_change 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 phishfort_plan_change. 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.
phishfort_plan_change is provided by the Phishfort MCP server (mychaelconnolly/phishfort-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 →