Begin a NEW workflow instance. Returns a server-minted { correlationId } that identifies this run. Call this once when the user starts a new task, then pass the returned correlationId to subsequent complete-slice / log-event-to-bus calls so a second task on the same connection stays isolated from...
AI agents invoke start-workflow to trigger actions in Mmc. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
This tool triggers execution of structured business processes with persistent state (correlationId tracking). While the immediate action is initialization rather than destructive, it commits to executing a workflow whose side effects depend on subsequent tool calls. The sequenced event bus architecture and correlation tracking indicate the tool orchestrates external operations, making it Execute rather than Write.
From the tool's definition 'Begin a NEW workflow instance' with 'server-minted { correlationId }' establishes and triggers external process execution; 'Call this once when the user starts a new task' indicates initiation of consequential operations whose effects depend on downstream…
Attacks that exploit this kind of access
Begin a NEW workflow instance. Returns a server-minted { correlationId } that identifies this run. Call this once when the user starts a new task, then pass the returned correlationId to subsequent complete-slice / log-event-to-bus calls so a second task on the same connection stays isolated from the first. The server is the sole minter — never invent a correlationId. It is categorised as a Execute tool in the Mmc MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Mmc MCP server in PolicyLayer and add a rule for start-workflow: 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 Mmc. Nothing to install.
start-workflow is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the start-workflow 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 start-workflow. 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.
start-workflow is provided by the Mmc MCP server (modelmycontext/mmc-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 →