AI agents call diff_against_disk to retrieve information from Kontexta without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and compares file content to identify differences, which is a non-destructive read operation. It has no side effects on the file system or data—it only reports on the state of files. The comparison itself does not modify, delete, or execute anything, making it a safe Read category with low severity.
From the tool's definition Tool name 'diff_against_disk' and description 'Diagnose drift between one file' indicate comparison/diagnosis of file contents without modification. The verb 'diagnose' and 'diff' are read-only operations that examine differences.
Attacks that exploit this kind of access
Diagnose drift between one file. It is categorised as a Read tool in the Kontexta MCP Server, which means it retrieves data without modifying state.
Register the Kontexta MCP server in PolicyLayer and add a rule for diff_against_disk: 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 Kontexta. Nothing to install.
diff_against_disk 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 diff_against_disk 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 diff_against_disk. 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.
diff_against_disk is provided by the Kontexta MCP server (safiyu/kontexta). 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 →