Manage OpenGraph custom node display configs and v9 extension schemas. info_type options: list - list all custom node configs get - get details for a specific node kind (needs: kind_name) create - create new node kind with display metadata (needs: custom_types_json) update - update a node kind's ...
AI agents use custom_nodes to create or update resources in BloodHound MCP Server — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your BloodHound MCP Server environment.
This tool spans Read (list, get), Write (create, update), and Destructive (delete) operations. Applying the most-severe-applicable rule, delete node kind could be irreversible, making Destructive the highest category. However, 'delete a node kind' refers to display configuration/metadata rather than actual AD attack path data, so the blast radius is more limited.
From the tool's definition create - create new node kind with display metadata, update - update a node kind's display config, delete - delete a node kind
Documented attack patterns abuse exactly the kind of access custom_nodes gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and BloodHound MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for custom_nodes:
{
"version": "1",
"default": "deny",
"tools": {
"custom_nodes": {
"limits": [
{
"counter": "custom_nodes_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}custom_nodes 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.
Manage OpenGraph custom node display configs and v9 extension schemas. info_type options: list - list all custom node configs get - get details for a specific node kind (needs: kind_name) create - create new node kind with display metadata (needs: custom_types_json) update - update a node kind's display config (needs: kind_name, config_json) delete - delete a node kind (needs: kind_name) validate_icon - validate icon config before creating/updating (needs: icon_config_json) extension_list - list OpenGraph extensions (BloodHound v9+) extension_upsert - create/update extension schema (needs: extension_json or extension_file_path) extension_delete - delete extension schema by ID (needs: extension_id) extension_edges - list extension edge kinds (optional: schemas, is_traversable) args: info_type: what to retrieve (default: list) kind_name: Custom node kind name (for get,update, delete) custom_types_json: JSON string or object for creating a new node kind (for create) config_json: JSON string or object for updating a node kind's display config (for update) icon_config_json: JSON string or object for validating icon config (for validate_icon) extension_json: JSON string or object for BloodHound v9 OpenGraph extension upsert extension_file_path: local JSON file path for BloodHound v9 OpenGraph extension upsert extension_id: OpenGraph extension ID for delete schemas: schema name or list of schema names for extension edge filtering is_traversable: bool or BloodHound filter string (for example: eq:true). It is categorised as a Write tool in the BloodHound MCP Server MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the BloodHound MCP Server MCP server in PolicyLayer and add a rule for custom_nodes: 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 BloodHound MCP Server. Nothing to install.
custom_nodes 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 custom_nodes 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 custom_nodes. 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.
custom_nodes is provided by the BloodHound MCP Server MCP server (mwnickerson/bloodhound_mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from BloodHound MCP Server, 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.
13 BloodHound MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.