AI agents use scf_create_evidence to create or update resources in Scf — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Scf environment.
This tool creates or registers a new evidence record in the compliance/security platform, modifying organizational state by adding a tracked item. This is a Write operation because it creates data that can be updated or removed, with side effects limited to the audit and compliance system.
From the tool's definition Tool description explicitly states "write — editor+ role" and "Create an evidence tracking record". The verb "create" and the phrase "starts tracking an evidence item" indicate reversible creation of a new data record.
Attacks that exploit this kind of access
Create an evidence tracking record from a catalog evidence ID (write — editor+ role). Starts tracking an evidence item for the organization. It is categorised as a Write tool in the Scf MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Scf MCP server in PolicyLayer and add a rule for scf_create_evidence: 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 Scf. Nothing to install.
scf_create_evidence 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 scf_create_evidence 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 scf_create_evidence. 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.
scf_create_evidence is provided by the Scf MCP server (mcp-server-scf). 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 →