Evaluation-only pre-write safety check. Verifies the AnchorID exists,
AI agents call guard_write to retrieve information from Anchord MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool is explicitly described as 'evaluation-only' and performs a 'safety check', meaning it reads and validates data without modifying anything. It checks that an AnchorID exists and evaluates proposed write operations without committing them. This is a read-class operation despite the 'write' in its name. Severity is low because misuse only affects decision-making, not data integrity directly.
From the tool's definition Evaluation-only pre-write safety check — the tool performs a read/validation operation only, verifying whether a proposed write is safe, without actually executing the write
Attacks that exploit this kind of access
Evaluation-only pre-write safety check. Verifies the AnchorID exists,. It is categorised as a Read tool in the Anchord MCP MCP Server, which means it retrieves data without modifying state.
Register the Anchord MCP server in PolicyLayer and add a rule for guard_write: 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 Anchord MCP. Nothing to install.
guard_write is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the guard_write 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 guard_write. 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.
guard_write is provided by the Anchord MCP server (nolenation04/anchord-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 →