Create an email draft. The draft is saved but NOT sent. The user must manually review and send it. When replying to a thread (in_reply_to_thread_id), recipients are auto-populated from the thread participants — the
AI agents use create_draft to create or update resources in AgentCloak — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your AgentCloak environment.
This tool creates or modifies email drafts, which is a Write operation. Although the draft is not sent, it persists in the system and represents irreversible creation of new data. Severity is medium because drafts are typically user-recoverable (can be deleted), but misuse could result in sensitive PII being drafted and left in the mailbox for discovery.
From the tool's definition Tool description states 'Create an email draft' and 'draft is saved' — this creates new data (a draft email message) in the email system.
Documented attack patterns abuse exactly the kind of access create_draft gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and AgentCloak, and nothing reaches the server without passing your rules. This is the rule we recommend for create_draft:
{
"version": "1",
"default": "deny",
"tools": {
"create_draft": {
"limits": [
{
"counter": "create_draft_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
} create_draft stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Create an email draft. The draft is saved but NOT sent. The user must manually review and send it. When replying to a thread (in_reply_to_thread_id), recipients are auto-populated from the thread participants — the. It is categorised as a Write tool in the AgentCloak MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the AgentCloak MCP server in PolicyLayer and add a rule for create_draft: 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 AgentCloak. Nothing to install.
create_draft is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the create_draft 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 create_draft. 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.
create_draft is provided by the AgentCloak MCP server (ryanfren/agentcloak). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from AgentCloak, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
8 AgentCloak tools catalogued and risk-classified — across an index of 43,000+ MCP servers.