Undo the last action in the connected score application.
AI agents use undo_last_action to create or update resources in Mcp Score — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Mcp Score environment.
This tool modifies the state of a score document by reversing a previous change. While it alters data, it is reversible (can be re-applied via redo) and does not permanently destroy information. It fits Write category as a modification operation.
From the tool's definition Tool name 'undo_last_action' and description 'Undo the last action in the connected score application' indicates it reverts a prior modification to the score.
Documented attack patterns abuse exactly the kind of access undo_last_action gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Mcp Score, and nothing reaches the server without passing your rules. This is the rule we recommend for undo_last_action:
{
"version": "1",
"default": "deny",
"tools": {
"undo_last_action": {
"limits": [
{
"counter": "undo_last_action_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}undo_last_action 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.
Undo the last action in the connected score application. It is categorised as a Write tool in the Mcp Score MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Mcp Score MCP server in PolicyLayer and add a rule for undo_last_action: 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 Mcp Score. Nothing to install.
undo_last_action 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 undo_last_action 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 undo_last_action. 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.
undo_last_action is provided by the Mcp Score MCP server (tskovlund/mcp-score). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Mcp Score, 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.
18 Mcp Score tools catalogued and risk-classified — across an index of 43,000+ MCP servers.