// SCAN REPORT

Rockfish MCP Server

48 tools. 22 can modify or destroy data without limits.

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

Last updated: 4 Jul 2026

22 can modify or destroy data

26 read-only

48 tools total

Community server · catalogue entry verified 04/07/2026

How to control Rockfish MCP Server ↓

TOOL MAP

What Rockfish MCP Server exposes to your agents

Read (26) Write / Execute (18) Destructive / Financial (4)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous Rockfish MCP Server tools

Destructive · High delete_database This tool permanently removes an entire database and all its contents. Destructive · High delete_dataset Deletion of datasets cannot be undone and represents permanent data loss. Destructive · High delete_model Deleting a model is a destructive operation that permanently removes a trained machine learning artifact. Destructive · High delete_worker_set This tool permanently removes a worker set, which cannot be undone.

22 of Rockfish MCP Server's 48 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Rockfish MCP Server

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

Deny destructive operations

{
  "delete_database": {
    "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

{
  "create_database": {
    "limits": [
      {
        "counter": "create_database_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "discover_schema": {
    "limits": [
      {
        "counter": "discover_schema_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 Rockfish MCP Server — 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 ROCKFISH →

Instant setup, no code required.

FULL CATALOGUE

All 48 Rockfish MCP Server tools

DESTRUCTIVE 4 tools

Destructive delete_database Delete a database Destructive delete_dataset Delete a dataset Destructive delete_model Delete a model Destructive delete_worker_set Delete a worker set

EXECUTE 7 tools

Execute create_workflow Create and run a new workflow Execute execute_nl_query Execute a natural language query over a dataset. Execute execute_query Execute an Analytics structured query over a dataset. Execute execute_sql_query Execute a SQL query against Rockfish datasets and return results in CSV format Execute inject_scenario Inject a synthetic anomaly scenario and optionally generate tests. Execute start_generation_workflow Start a workflow to generate synthetic data from a trained TabGAN model. Execute start_training_workflow Build and start a training workflow using a cached config. Automatically detects model type from train_config

WRITE 11 tools

Write create_database Create a new database Write create_dataset Create a new dataset Write create_project Create a new project Write create_worker_set Create a new worker set Write generate_test_suite Generate Q&A test cases for a dataset from ID or raw CSV. Write update_database Update a database Write update_dataset Update a dataset Write update_project Update a project Write update_train_config Update training configuration. Modify model hyperparameters (epochs, batch_size, etc.) and/or encoder field cl Write update_workflow Update a workflow Write upload_model Upload a new model

READ 26 tools

Read discover_schema Analyze dataset structure, classify columns, and compute measurement stats. Read get_active_organization Get active organization Read get_active_project Get active project Read get_database Get a specific database by ID Read get_dataset Get a specific dataset by ID Read get_dataset_schema Get the schema for a dataset Read get_marginal_distribution_score Calculate marginal distribution score between two datasets (real vs synthetic). Lower score means more similar Read get_model Get a specific model by ID Read get_project Get a specific project by ID Read get_trained_model_id Extract the trained model ID from a completed training workflow. Only works on COMPLETED or FINALIZED workflow Read get_worker_set Get a specific worker set by ID Read get_worker_set_actions List the actions for a worker-set Read get_workflow Get a specific workflow by ID Read get_workflow_logs Stream logs from a workflow execution with configurable log level and collection timeout. Collects logs for sp Read list_available_actions List actions (or workers) from all worker-sets Read list_databases List all databases Read list_datasets List all datasets Read list_models List all models Read list_organizations List all organizations Read list_projects List all projects Read list_worker_groups List all worker groups Read list_worker_sets List all worker sets Read list_workflows List all workflows Read obtain_synthetic_dataset_id Extract the synthetic dataset ID from a completed generation workflow. Only works on COMPLETED or FINALIZED wo Read obtain_train_config Generate training configuration for synthetic data models. Returns a cached train_config_id for use in start_t Read plot_distribution Generate distribution plot comparing datasets for a specific column. Returns base64-encoded PNG image. Automat

EXPLORE

Related servers

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

FAQ

Questions about Rockfish MCP Server

Can an AI agent delete data through the Rockfish MCP Server MCP server? +

Yes. The Rockfish MCP Server server exposes 4 destructive tools including delete_database, delete_dataset, delete_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 Rockfish MCP Server? +

The Rockfish MCP Server server has 11 write tools including create_database, create_dataset, create_project. 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 Rockfish MCP Server.

How many tools does the Rockfish MCP Server MCP server expose? +

48 tools across 4 categories: Destructive, Execute, Read, Write. 26 are read-only. 22 can modify, create, or delete data.

How do I enforce a policy on Rockfish MCP Server? +

Register the Rockfish MCP Server 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 Rockfish MCP Server tool call.

Deterministic rules across all 48 Rockfish MCP Server tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON ROCKFISH →

Instant setup, no code required.

48 Rockfish MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.