AI agents call check_downloads to retrieve information from Grounded without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool queries and returns package download statistics from npm or PyPI. It is a pure read operation that retrieves existing data without modifying, deleting, or executing anything. The blast radius is minimal — misuse would only return statistical information already public. No financial, destructive, or code execution implications exist.
From the tool's definition Tool description states 'Check the weekly download count' — a retrieval operation with no side effects. The server description confirms it 'enables users to perform technical lookups' and 'provides real-time data for package versions, download counts, and…
Attacks that exploit this kind of access
Check the weekly download count of a package on npm or PyPI. It is categorised as a Read tool in the Grounded MCP Server, which means it retrieves data without modifying state.
Register the Grounded MCP server in PolicyLayer and add a rule for check_downloads: 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 Grounded. Nothing to install.
check_downloads 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 check_downloads 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 check_downloads. 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.
check_downloads is provided by the Grounded MCP server (nghizas/grounded-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 →