Take analyze_spec_quality output and produce a PM-facing
AI agents call propose_spec_improvements to retrieve information from Mk Spec Master without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool appears to consume analysis output and produce a proposal or report directed at a PM. Generating recommendations is a read/report operation with no side effects. The description is incomplete (cut off), which lowers confidence slightly, but the phrasing 'produce a PM-facing' strongly implies a report/document output rather than a mutation.
From the tool's definition 'propose_spec_improvements' takes 'analyze_spec_quality output and produce a PM-facing' suggestion — it generates recommendations/proposals without writing or modifying any data
Attacks that exploit this kind of access
Take analyze_spec_quality output and produce a PM-facing. It is categorised as a Read tool in the Mk Spec Master MCP Server, which means it retrieves data without modifying state.
Register the Mk Spec Master MCP server in PolicyLayer and add a rule for propose_spec_improvements: 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 Mk Spec Master. Nothing to install.
propose_spec_improvements 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 propose_spec_improvements 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 propose_spec_improvements. 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.
propose_spec_improvements is provided by the Mk Spec Master MCP server (kao273183/mk-spec-master). 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 →