评审 proposal 或 design 文档，识别潜在问题（完整性、可行性、安全、边界条件、清晰度）
AI agents call openspec_critique_proposal to retrieve information from OpenSpec MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool examines and evaluates existing documents to surface issues, which is fundamentally a retrieval and analysis operation with no side effects. It does not create, modify, delete, execute code, or move funds. While it provides security-related analysis, that analysis itself does not trigger security changes—it only identifies concerns for human review.
From the tool's definition Tool performs critique/review of proposals and design documents, identifying potential issues. The description indicates analysis and identification of problems (completeness, feasibility, security, edge cases, clarity) without modifying or executing anything.
Documented attack patterns abuse exactly the kind of access openspec_critique_proposal gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and OpenSpec MCP, and nothing reaches the server without passing your rules. This is the rule we recommend for openspec_critique_proposal:
{
"version": "1",
"default": "deny",
"tools": {
"openspec_critique_proposal": {}
}
}openspec_critique_proposal is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
评审 proposal 或 design 文档，识别潜在问题（完整性、可行性、安全、边界条件、清晰度）. It is categorised as a Read tool in the OpenSpec MCP MCP Server, which means it retrieves data without modifying state.
Register the OpenSpec MCP server in PolicyLayer and add a rule for openspec_critique_proposal: 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 OpenSpec MCP. Nothing to install.
openspec_critique_proposal 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 openspec_critique_proposal 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 openspec_critique_proposal. 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.
openspec_critique_proposal is provided by the OpenSpec MCP server (lumiaqian/openspec-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from OpenSpec MCP, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
40 OpenSpec MCP tools catalogued and risk-classified — across an index of 43,000+ MCP servers.