list buckets/objects, get/put/delete objects.
AI agents call S3 to permanently remove resources in Aws Claw — typically in cleanup and lifecycle workflows. It does its job in a single call, and there is no undo.
S3 object deletion is irreversible and can cause significant data loss. An AI agent with access to this tool could permanently destroy critical business data, backups, or compliance records. The tool's access to multiple buckets via 'list buckets' amplifies the blast radius. This is classified as Destructive rather than Execute because the primary risk is data destruction, not arbitrary code execution.
From the tool's definition Tool explicitly supports 'delete objects' which irreversibly removes data from S3 buckets. While it also supports read (list/get) and write (put) operations, the delete capability makes this destructive.
Attacks that exploit this kind of access
list buckets/objects, get/put/delete objects. It is categorised as a Destructive tool in the Aws Claw MCP Server, which means it can permanently delete or destroy data. Block by default and require explicit approval.
Register the Aws Claw MCP server in PolicyLayer and add a rule for S3: 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 Aws Claw. Nothing to install.
S3 is a Destructive tool with critical risk. Critical-risk tools should be blocked by default and only enabled with explicit human approval.
Yes. Add a rate_limit block to the S3 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 S3. 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.
S3 is provided by the Aws Claw MCP server (necatiarslan/awsclaw). 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 →