18 tools. 9 can modify or destroy data without limits.
3 destructive tools with no built-in limits. Policy required.
Last updated:
Community server · catalogue entry verified 30/06/2026
Running more than one server? Check your whole stack →
9 of Litellm Admin's 18 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.
PolicyLayer is an MCP gateway — it sits between your AI agents and Litellm Admin, and nothing reaches the server without passing your rules. These are the rules we recommend:
{
"litellm_key_delete": {
"deny_if": [
{
"conditions": [],
"on_deny": "Blocked by default. Requires approval."
}
]
}
} Destructive tools should never be available to autonomous agents without human approval.
{
"litellm_key_generate": {
"limits": [
{
"counter": "litellm_key_generate_per_hour",
"window": "hour",
"max": 30,
"scope": "grant"
}
]
}
} Prevents bulk unintended modifications from agents caught in loops.
{
"litellm_key_info": {
"limits": [
{
"counter": "litellm_key_info_per_minute",
"window": "minute",
"max": 60,
"scope": "grant"
}
]
}
} Controls API costs and prevents retry loops from exhausting upstream rate limits.
Instant setup, no code required.
Other MCP servers with similar tools — same risk classification, starter policies for each.
Yes. The Litellm Admin server exposes 3 destructive tools including litellm_key_delete, litellm_model_delete, litellm_user_delete. These permanently remove resources with no undo. PolicyLayer blocks destructive tools by default so they never reach the upstream server.
The Litellm Admin server has 6 write tools including litellm_key_generate, litellm_key_update, litellm_model_create. Set a rate limit in your policy -- for example, 10 calls per hour prevents an agent from making more than 10 modifications per hour. PolicyLayer enforces this at the gateway, before calls reach Litellm Admin.
18 tools across 3 categories: Destructive, Read, Write. 9 are read-only. 9 can modify, create, or delete data.
Register the Litellm Admin MCP server in PolicyLayer, apply the suggested rules above (adjust the limits to your use case), and point your AI client at the PolicyLayer proxy URL instead of the server directly. Your agents keep the same tools; PolicyLayer evaluates every call against policy before it executes. Nothing to install, live in minutes.
Deterministic rules across all 18 Litellm Admin tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Instant setup, no code required.
18 Litellm Admin tools catalogued and risk-classified — across an index of 43,000+ MCP servers.
These policies come from Litellm Admin's registry record.
The record behind this page: verified identity, auth posture, risk grade, every tool classified, recommended policy — re-checked continuously.
Teams ship this data inside their own products. See what a licence covers →