AI agents use learn to create or update resources in Project Tessera — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Project Tessera environment.
The tool writes new knowledge/data into the local workspace memory system by saving and indexing it. This is a reversible modification (data can be updated or removed later), making it Write rather than Read (which would only retrieve) or Destructive (which would permanently delete).
From the tool's definition Tool description states 'Save and immediately index new knowledge' — this creates new data entries in the vector store without deletion or financial impact.
Documented attack patterns abuse exactly the kind of access learn gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Project Tessera, and nothing reaches the server without passing your rules. This is the rule we recommend for learn:
{
"version": "1",
"default": "deny",
"tools": {
"learn": {
"limits": [
{
"counter": "learn_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}learn 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.
Save and immediately index new knowledge. It is categorised as a Write tool in the Project Tessera MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Project Tessera MCP server in PolicyLayer and add a rule for learn: 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 Project Tessera. Nothing to install.
learn 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 learn 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 learn. 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.
learn is provided by the Project Tessera MCP server (besslframework-stack/project-tessera). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Project Tessera, 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.
43 Project Tessera tools catalogued and risk-classified — across an index of 43,000+ MCP servers.