AI agents use papercortex_classify to create or update resources in PaperCortex — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your PaperCortex environment.
Classification assigns metadata/tags/categories to a document, which modifies the document's attributes in Paperless-ngx. This is a reversible write operation (classification can be changed), not merely a read. The blast radius is medium — misclassification could misfiled documents but is correctable.
From the tool's definition Auto-classify a document using local AI
Documented attack patterns abuse exactly the kind of access papercortex_classify gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and PaperCortex, and nothing reaches the server without passing your rules. This is the rule we recommend for papercortex_classify:
{
"version": "1",
"default": "deny",
"tools": {
"papercortex_classify": {
"limits": [
{
"counter": "papercortex_classify_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}papercortex_classify 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.
Auto-classify a document using local AI. It is categorised as a Write tool in the PaperCortex MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the PaperCortex MCP server in PolicyLayer and add a rule for papercortex_classify: 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 PaperCortex. Nothing to install.
papercortex_classify 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 papercortex_classify 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 papercortex_classify. 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.
papercortex_classify is provided by the PaperCortex MCP server (renefichtmueller/papercortex). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from PaperCortex, 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.
5 PaperCortex tools catalogued and risk-classified — across an index of 43,000+ MCP servers.