Copy a template ability to a character
AI agents use learn_ability to create or update resources in DMCP — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your DMCP environment.
This tool creates/adds new data (an ability) to a character record, which is a reversible write operation within a game state management system. It does not delete data, execute code, or involve financial transactions. The blast radius is low as it only affects in-game character state in a text-based RPG context.
From the tool's definition Copy a template ability to a character
Documented attack patterns abuse exactly the kind of access learn_ability gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and DMCP, and nothing reaches the server without passing your rules. This is the rule we recommend for learn_ability:
{
"version": "1",
"default": "deny",
"tools": {
"learn_ability": {
"limits": [
{
"counter": "learn_ability_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}learn_ability 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.
Copy a template ability to a character. It is categorised as a Write tool in the DMCP MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the D MCP server in PolicyLayer and add a rule for learn_ability: 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 DMCP. Nothing to install.
learn_ability 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_ability 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_ability. 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_ability is provided by the D MCP server (shawnrushefsky/dmcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from DMCP, 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.
204 DMCP tools catalogued and risk-classified — across an index of 43,000+ MCP servers.