AI agents use set_color to create or update resources in Paint MCP — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Paint MCP environment.
This tool creates or modifies data (the active drawing color) reversibly. It does not retrieve data (Read), execute arbitrary code (Execute), delete anything (Destructive), or involve financial operations (Financial). The severity is low because changing a color setting has minimal blast radius—any unwanted color choices can be trivially undone by calling set_color again with different values.
From the tool's definition set_color modifies the canvas state by changing the drawing color parameter (RGB values 0-255). The description explicitly states it 'set[s] the drawing color', which is a reversible modification to the painter's configuration state rather than a permanent…
Attacks that exploit this kind of access
Set the drawing color (RGB, each 0-255). It is categorised as a Write tool in the Paint MCP MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Paint MCP server in PolicyLayer and add a rule for set_color: 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 Paint MCP. Nothing to install.
set_color 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 set_color 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 set_color. 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.
set_color is provided by the Paint MCP server (joeyballentine/paint-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 →