Undelivered notifications for OpenClaw's heartbeat to surface, then mark read.
AI agents call pending_notifications to retrieve information from surveyHelper without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves notification data and performs a reversible read-status update. It is a pure read operation with a minor state mutation (marking as read). The scope is limited to notification metadata on a local machine (per server description), posing minimal risk if misused by an AI agent.
From the tool's definition Tool description states it surfaces undelivered notifications and marks them read. The verb 'surface' indicates retrieval/querying of existing notifications, and 'mark read' is a metadata state change (not destructive deletion).
Attacks that exploit this kind of access
Undelivered notifications for OpenClaw's heartbeat to surface, then mark read. It is categorised as a Read tool in the surveyHelper MCP Server, which means it retrieves data without modifying state.
Register the surveyHelper MCP server in PolicyLayer and add a rule for pending_notifications: 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 surveyHelper. Nothing to install.
pending_notifications 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 pending_notifications 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 pending_notifications. 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.
pending_notifications is provided by the surveyHelper MCP server (rich7420/surveyhelper). 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 →