AI agents use write-query to create or update resources in Mcp Libsql — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Mcp Libsql environment.
The tool modifies data through INSERT, UPDATE, and DELETE statements. While DELETE can remove data, the presence of transaction support means deletions can be rolled back, classifying this as Write rather than Destructive. The high severity reflects the risk of an AI agent inadvertently modifying or removing data at scale, but the reversibility through transactions prevents critical classification.
From the tool's definition Tool description explicitly states 'INSERT/UPDATE/DELETE operations with transaction support'. INSERT and UPDATE are write operations that modify data reversibly; DELETE is listed but within a write-query tool (not a dedicated destructive tool), and…
Documented attack patterns abuse exactly the kind of access write-query gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Mcp Libsql, and nothing reaches the server without passing your rules. This is the rule we recommend for write-query:
{
"version": "1",
"default": "deny",
"tools": {
"write-query": {
"limits": [
{
"counter": "write-query_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}write-query 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.
INSERT/UPDATE/DELETE operations with transaction support. It is categorised as a Write tool in the Mcp Libsql MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Mcp Libsql MCP server in PolicyLayer and add a rule for write-query: 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 Libsql. Nothing to install.
write-query 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 write-query 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 write-query. 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.
write-query is provided by the Mcp Libsql MCP server (xexr/mcp-libsql). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Mcp Libsql, 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.
6 Mcp Libsql tools catalogued and risk-classified — across an index of 43,000+ MCP servers.