Replace a value in the cache only if the key exists. Args: key: The key to replace value: The new value expire: Optional expiration time in seconds Returns: Success message or error message
AI agents use cache_replace to create or update resources in AWS — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your AWS environment.
This tool creates or modifies data reversibly by replacing cached values. It does not delete data (hence not Destructive), does not execute arbitrary code (hence not Execute), does not move money (hence not Financial), and does not merely read data (hence not Read).
From the tool's definition Tool modifies cached data via replace operation on an existing key. Description states 'Replace a value in the cache' which is a write/update operation.
Documented attack patterns abuse exactly the kind of access cache_replace gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and AWS, and nothing reaches the server without passing your rules. This is the rule we recommend for cache_replace:
{
"version": "1",
"default": "deny",
"tools": {
"cache_replace": {
"limits": [
{
"counter": "cache_replace_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}cache_replace stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Replace a value in the cache only if the key exists. Args: key: The key to replace value: The new value expire: Optional expiration time in seconds Returns: Success message or error message. It is categorised as a Write tool in the AWS MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the AWS MCP server in PolicyLayer and add a rule for cache_replace: 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. Nothing to install.
cache_replace is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the cache_replace 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 cache_replace. 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.
cache_replace is provided by the AWS MCP server (@awslabs/mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from AWS, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
300 AWS tools catalogued and risk-classified — across an index of 43,000+ MCP servers.