AI agents use ksef_start_invoice_draft to create or update resources in KSeF MCP — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your KSeF MCP environment.
The tool creates a new invoice draft object in the system, which is a reversible write operation (drafts can typically be discarded). It does not submit or finalise the invoice, so it does not reach Execute or Financial severity, but misuse could lead to cluttered or fraudulent draft records, warranting medium severity.
From the tool's definition 'Create an invoice draft' — creates a new draft resource; returning a checklist is a read side-effect of the creation action.
Attacks that exploit this kind of access
Create an invoice draft and return the missing-field checklist. It is categorised as a Write tool in the KSeF MCP MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the KSeF MCP server in PolicyLayer and add a rule for ksef_start_invoice_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 KSeF MCP. Nothing to install.
ksef_start_invoice_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 ksef_start_invoice_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 ksef_start_invoice_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.
ksef_start_invoice_draft is provided by the KSeF MCP server (olegtyshcneko/ksef-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 →