AI agents invoke compile to trigger actions in McFlow. 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.
The 'compile' action with code/prompt injection capability is an Execute risk because it runs or processes external code within the workflow context. While not Destructive (no irreversible deletion), it goes beyond Write (mere modification) by actively executing code whose behavior depends on the injected content.
From the tool's definition Tool description states 'Compile all workflows by injecting external code/prompt files' — this involves executing code (injected code/prompts) into workflows, which triggers external operations whose effects depend on what code is injected.
Attacks that exploit this kind of access
Compile all workflows by injecting external code/prompt files. It is categorised as a Execute tool in the McFlow MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the McFlow MCP server in PolicyLayer and add a rule for compile: 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 McFlow. Nothing to install.
compile 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 compile 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 compile. 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.
compile is provided by the McFlow MCP server (mckinleymedia/mcflow-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 →