// SCAN REPORT

Neo4j Gds

72 tools. 19 can modify or destroy data without limits.

3 destructive tools with no built-in limits. Policy required.

Last updated: 27 Jun 2026

19 can modify or destroy data

53 read-only

72 tools total

Community server · catalogue entry verified 27/06/2026

How to control Neo4j Gds ↓

TOOL MAP

What Neo4j Gds exposes to your agents

Read (53) Write / Execute (16) Destructive / Financial (3)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous Neo4j Gds tools

Destructive · High delete_session Delete operations are irreversible and destructive by definition. Destructive · High drop_graph The 'drop' operation in database systems is a classic destructive action that cannot be undone without restoration from backups. Destructive · High drop_model This tool permanently deletes trained ML models from the GDS catalog. Execute · Medium CELF CELF (Cost-Effective Lazy Forward) is a known graph algorithm for influence maximization.

19 of Neo4j Gds's 72 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Neo4j Gds

PolicyLayer is an MCP gateway — it sits between your AI agents and Neo4j Gds, and nothing reaches the server without passing your rules. These are the rules we recommend:

Deny destructive operations

{
  "delete_session": {
    "deny_if": [
      {
        "conditions": [],
        "on_deny": "Blocked by default. Requires approval."
      }
    ]
  }
}

Destructive tools should never be available to autonomous agents without human approval.

Rate limit write operations

{
  "graph_sage_train": {
    "limits": [
      {
        "counter": "graph_sage_train_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "a_star_shortest_path": {
    "limits": [
      {
        "counter": "a_star_shortest_path_per_minute",
        "window": "minute",
        "max": 60,
        "scope": "grant"
      }
    ]
  }
}

Controls API costs and prevents retry loops from exhausting upstream rate limits.

Create a free account and register Neo4j Gds — nothing to install.
Add these rules — paste them, or build them visually. Tune the limits to your setup.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

ENFORCE POLICY ON NEO4J GDS →

Instant setup, no code required.

FULL CATALOGUE

All 72 Neo4j Gds tools

DESTRUCTIVE 3 tools

Destructive delete_session delete_session Destructive drop_graph drop_graph Destructive drop_model Remove a trained machine learning model from the GDS model catalog to free memory.

EXECUTE 11 tools

Execute CELF CELF Execute graph_sage_predict Generates node embeddings using a previously trained GraphSAGE model from the model catalog (see graph_sage_tr Execute HDBSCAN Perform HDBSCAN clustering on the graph. HDBSCAN, which stands for Hierarchical Density-Based Spatial Clusteri Execute HITS HITS Execute k_means_clustering K-Means clustering is an unsupervised learning algorithm that is used to solve clustering problems. Execute node2vec Node2Vec is a node embedding algorithm that computes a vector representation of a node based on second-order r Execute pagerank pagerank Execute predict_link_prediction Predict new relationships in a projected graph using a trained link prediction model from the model catalog Execute predict_node_regression Predict numeric values for nodes in a projected graph using a trained node regression model from the model cat Execute project_graph_cypher project_graph_cypher Execute speaker_listener_label_propagation The Speaker-Listener Label Propagation Algorithm (SLLPA) is a variation of the Label Propagation algorithm tha

WRITE 5 tools

Write graph_sage_train Trains a GraphSAGE model and stores it in the GDS model catalog under modelName. Write k_nearest_neighbors The K-Nearest Neighbors algorithm computes a distance value for all node pairs in the graph and creates new re Write prize_collecting_steiner_tree A spanning tree is a graph such that there is exactly one path between any two nodes in the set. A graph can h Write train_link_prediction_model Train a link prediction model on a projected graph and store it in the GDS model catalog under modelName. Write train_node_regression_model Train a node regression model on a projected graph and store it in the GDS model catalog under modelName.

READ 53 tools

Read a_star_shortest_path The A (pronounced Read all_pairs_shortest_paths The All Pairs Shortest Path (APSP) calculates the shortest (weighted) path between all pairs of nodes. Read approximate_maximum_k_cut A k-cut of a graph is an assignment of its nodes into k disjoint communities. Read article_rank article_rank Read articulation_points Find all the articulation points. Given a graph, an articulation point is a node whose removal increases the n Read bellman_ford_single_source_shortest_path The Bellman-Ford Path algorithm computes the shortest path between nodes. Read betweenness_centrality betweenness_centrality Read breadth_first_search The Breadth First Search algorithm is a graph traversal algorithm that given a start node visits nodes in orde Read bridges bridges Read closeness_centrality closeness_centrality Read conductance conductance Read count_nodes count_nodes Read degree_centrality degree_centrality Read delta_stepping_shortest_path The Delta-Stepping Shortest Path algorithm computes all shortest paths between a source node and all reachable Read depth_first_search The Depth First Search algorithm is a graph traversal that starts at a given node and explores as far as possi Read dijkstra_single_source_shortest_path The Dijkstra Shortest Path algorithm computes the shortest path between nodes. Read eigenvector_centrality eigenvector_centrality Read fast_rp Fast Random Projection (FastRP) is a node embedding algorithm that computes a vector representation for every Read find_shortest_path Find the shortest path between two nodes using Dijkstra Read get_graph_info get_graph_info Read get_node_labels get_node_labels Read get_node_properties_keys get_node_properties_keys Read get_relationship_properties_keys get_relationship_properties_keys Read get_relationship_types get_relationship_types Read harmonic_centrality harmonic_centrality Read hashgnn HashGNN is a node embedding algorithm which resembles Graph Neural Networks but does not include a model or re Read k_1_coloring The K-1 Coloring algorithm assigns a color to every node in the graph, Read k_core_decomposition The K-core decomposition constitutes a process of separates the nodes in a graph into groups based on the degr Read label_propagation The Label Propagation algorithm (LPA) is a fast algorithm for finding communities in a graph. Read leiden The Leiden algorithm is an algorithm for detecting communities in large networks. Read list_graphs list_graphs Read list_models List all machine learning models in the GDS model catalog (GraphSAGE, node classification, link prediction, no Read list_sessions list_sessions Read local_clustering_coefficient The Local Clustering Coefficient algorithm computes the local clustering coefficient for each node in the grap Read longest_path Finding the longest path that leads to a node in a graph is possible to do in linear time for the special case Read louvain The Louvain method is an algorithm to detect communities in large networks. Read max_flow Given source nodes, target nodes and relationships with capacity constraints, the max-flow algorithm assigns a Read minimum_directed_steiner_tree Given a source node and a list of target nodes, a directed spanning tree in which there exists a path from the Read minimum_weight_spanning_tree The Minimum Weight Spanning Tree (MST) starts from a given node, finds all its reachable nodes and returns the Read modularity_metric Modularity is a metric that allows you to evaluate the quality of a community detection. Read modularity_optimization The Modularity Optimization algorithm tries to detect communities in the graph based on their modularity. Read node_similarity The Node Similarity algorithm compares a set of nodes based on the nodes they are connected to. Read predict_node_classification Predict class labels for nodes in a projected graph using a trained node classification model from the model c Read random_walk Random Walk is an algorithm that provides random paths in a graph. Read recreate_session recreate_session Read stream_node_properties stream_node_properties Read stream_relationship_properties stream_relationship_properties Read stream_relationships stream_relationships Read strongly_connected_components The Strongly Connected Components (SCC) algorithm finds maximal sets of connected nodes in a directed graph. Read train_node_classification_model Train a node classification model on a projected graph and store it in the GDS model catalog under modelName. Read triangle_count The Triangle Count algorithm counts the number of triangles for each node in the graph. Read weakly_connected_components The Weakly Connected Components (WCC) algorithm finds sets of connected nodes in directed and undirected graph Read yens_shortest_paths Yen

EXPLORE

Related servers

Other MCP servers with similar tools — same risk classification, starter policies for each.

FAQ

Questions about Neo4j Gds

Can an AI agent delete data through the Neo4j Gds MCP server? +

Yes. The Neo4j Gds server exposes 3 destructive tools including delete_session, drop_graph, drop_model. These permanently remove resources with no undo. PolicyLayer blocks destructive tools by default so they never reach the upstream server.

How do I prevent bulk modifications through Neo4j Gds? +

The Neo4j Gds server has 5 write tools including graph_sage_train, k_nearest_neighbors, prize_collecting_steiner_tree. 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 Neo4j Gds.

How many tools does the Neo4j Gds MCP server expose? +

72 tools across 3 categories: Destructive, Read, Write. 53 are read-only. 19 can modify, create, or delete data.

How do I enforce a policy on Neo4j Gds? +

Register the Neo4j Gds 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.

Enforce policy on every Neo4j Gds tool call.

Deterministic rules across all 72 Neo4j Gds tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON NEO4J GDS →

Instant setup, no code required.

72 Neo4j Gds tools catalogued and risk-classified — across an index of 43,000+ MCP servers.