WATCHER: get a signed callback when a new academic paper matching your query is published (arXiv / PubMed / Semantic Scholar). Arm once, pay once. Pass query (keywords, author, topic). Fires once per new paper (deduped by source id); bounded by maxFires/expiry. Existing results at arm time are ba...
AI agents call watchers.paper to retrieve information from Mcp without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This is fundamentally a read operation that queries public academic paper databases and sets up notifications. While it establishes a persistent watcher, it only retrieves and filters publicly available information without creating, modifying, or destroying data.
From the tool's definition The tool retrieves and monitors academic papers matching a query from arXiv, PubMed, and Semantic Scholar. It 'gets a signed callback when a new academic paper matching your query is published' and 'returns a watcherId'.
Attacks that exploit this kind of access
WATCHER: get a signed callback when a new academic paper matching your query is published (arXiv / PubMed / Semantic Scholar). Arm once, pay once. Pass query (keywords, author, topic). Fires once per new paper (deduped by source id); bounded by maxFires/expiry. Existing results at arm time are baselined. Signed (verify offline) + retried; recoverable via watchers.status. Returns a watcherId. It is categorised as a Read tool in the Mcp MCP Server, which means it retrieves data without modifying state.
Register the MCP server in PolicyLayer and add a rule for watchers.paper: 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 Mcp. Nothing to install.
watchers.paper 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 watchers.paper 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 watchers.paper. 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.
watchers.paper is provided by the MCP server (@2sio/mcp). 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 →