AI agents use mark_invoice_sent to create or update resources in Sevdesk — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Sevdesk environment.
This tool changes invoice status, which is a write operation that modifies data. While it affects financial documents, it does not create financial obligations, move money, or perform irreversible deletions—those would be Financial or Destructive. The state transition itself is reversible (an invoice can be marked unsent or have its status changed back), making it Write rather than Destructive.
From the tool's definition Transitions invoice status from draft to open state in sevdesk accounting system. The tool modifies invoice state ('mark_invoice_sent') which is a reversible state change in an accounting context.
Attacks that exploit this kind of access
Mark invoice as sent in sevdesk v2.0 (transitions draft→open). It is categorised as a Write tool in the Sevdesk MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Sevdesk MCP server in PolicyLayer and add a rule for mark_invoice_sent: 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 Sevdesk. Nothing to install.
mark_invoice_sent 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 mark_invoice_sent 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 mark_invoice_sent. 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.
mark_invoice_sent is provided by the Sevdesk MCP server (codestra/sevdesk-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 →