AI agents call get_inbox to retrieve information from Reddirect without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool queries existing data from the user's inbox without any side effects. It does not create, modify, delete, or execute operations. While inbox messages may contain personal information, the risk is limited to information disclosure to an authorized user accessing their own account, consistent with Read category tools.
From the tool's definition Tool retrieves inbox messages (unread, mentions, comment replies) with no modification capability indicated. The verb 'get' and description 'Get your Reddit inbox messages' clearly denotes data retrieval only.
Attacks that exploit this kind of access
Get your Reddit inbox messages including unread, mentions, and comment replies. It is categorised as a Read tool in the Reddirect MCP Server, which means it retrieves data without modifying state.
Register the Reddirect MCP server in PolicyLayer and add a rule for get_inbox: 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 Reddirect. Nothing to install.
get_inbox 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 get_inbox 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 get_inbox. 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.
get_inbox is provided by the Reddirect MCP server (jeebus87/reddirect). 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 →