// SCAN REPORT

GraphHub

32 tools. 6 can modify or destroy data without limits.

1 destructive tool with no built-in limits. Policy required.

Last updated: 11 Jun 2026

6 can modify or destroy data

26 read-only

32 tools total

Community server · catalogue entry verified 11/06/2026

How to control GraphHub ↓

TOOL MAP

What GraphHub exposes to your agents

Read (26) Write / Execute (5) Destructive / Financial (1)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous GraphHub tools

Destructive · High forget This tool irreversibly removes data (observations) from the GraphHub knowledge graph memory system. Execute · High query_graph Executing arbitrary Cypher queries against a graph database is an Execute-level action. Write · Low remember The 'remember' tool creates or modifies persistent session memory entries, which is characteristic of Write operations. Write · Medium update_observation This is a Write operation because it creates or modifies data reversibly.

6 of GraphHub's 32 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control GraphHub

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

Deny destructive operations

{
  "forget": {
    "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

{
  "remember_bugfix": {
    "limits": [
      {
        "counter": "remember_bugfix_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "batch_context": {
    "limits": [
      {
        "counter": "batch_context_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 GraphHub — 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 GRAPHHUB →

Free to start. No card required.

FULL CATALOGUE

All 32 GraphHub tools

DESTRUCTIVE 1 tools

Destructive forget Delete observations from memory. Can delete by ID, session, project, type, or date.

EXECUTE 1 tools

Execute query_graph Run a direct Cypher query against the codebase graph database. Use this to find complex relationships.

WRITE 4 tools

Write remember_bugfix Record a resolved bug so future agents can reuse the fix pattern. Stores symptom, root cause, and fix as a str Write remember_skill_choice Cache which .skill.md / SKILL.md file was used for a task so similar future tasks route to the same skill with Write remember Save a learning, decision, finding, or other observation to persistent memory. Searchable via recall with sema Write update_observation Update an existing observation

READ 26 tools

Read batch_context Get definition + caller/callee counts for multiple symbols in one call. Replaces N get_context calls when an a Read changed_symbols List the symbols in files changed by the current branch (or working tree / staged diff). Each entry includes a Read check_arch_rules Enforce architecture boundary rules — detect files that import across forbidden layer boundaries. Rules can be Read explain_search Explain how a search query will be processed. Shows tokenization and search strategy. Read find_cycles Detect circular import chains or mutual-recursion call cycles in the codebase. Returns each cycle as an ordere Read find_dead_code Find functions, methods, and classes that are never called by anything in the graph. Useful before refactors o Read find_duplicates Find functions that are semantically similar to a given symbol — likely duplicate or near-duplicate implementa Read find_similar Find symbols that are semantically similar to a given symbol. Useful for finding related code. Read find_tech_debt Surface TODO/FIXME/HACK/OPTIMIZE markers stored in the graph during indexing. Returns symbols sorted by risk s Read get_context Get all callers and callees of a symbol by name. Shows the full call graph context around a function or class. Read get_file_symbols Retrieve all symbols (classes, functions, etc.) defined in a specific file. Read get_hierarchy Explore the inheritance and interface implementation tree for a class or interface. Returns ancestors (parent Read get_observation Get a specific observation by its ID. Read get_test_coverage Show which functions and methods in the codebase are exercised by test files. Identifies uncovered symbols so Read impact_analysis Analyze all symbols that directly call or import a given symbol. Shows what would break if the symbol changes. Read memory_stats Get statistics about stored observations: totals, counts by type, by project, by importance. Read recall Search memory for past observations using natural language. Returns results ranked by semantic similarity. Read recall_bugfix Find past bug fixes with similar symptoms. Returns structured fix patterns (symptom/root_cause/fix) ranked by Read recall_skill_choice Find which skill was used for similar past tasks. Skip the skill-selection decision for common task patterns. Read related_observations Get all observations that are linked to a specific code symbol. Read review_diff Generate a structured pre-merge review of the current git diff. Returns each changed symbol with its blast rad Read search_by_name Search for symbols by their name. Supports exact or fuzzy matching. Read search_grouped Search for code and group results by file. Useful for understanding which files are most relevant. Read semantic_search Search for code using natural language. Supports semantic (meaning-based), keyword, or hybrid search modes. Read timeline View observations chronologically. Useful for reviewing what happened during a session or project. Read debug_trace One-shot debugging entry point. Given a bug description or error message, returns ranked candidate symbols enr

EXPLORE

Related servers

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

FAQ

Questions about GraphHub

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

Yes. The GraphHub server exposes 1 destructive tools including forget. 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 GraphHub? +

The GraphHub server has 4 write tools including remember_bugfix, remember_skill_choice, remember. 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 GraphHub.

How many tools does the GraphHub MCP server expose? +

32 tools across 3 categories: Destructive, Read, Write. 26 are read-only. 6 can modify, create, or delete data.

How do I enforce a policy on GraphHub? +

Register the GraphHub 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 GraphHub tool call.

Deterministic rules across all 32 GraphHub tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON GRAPHHUB →

Free to start. No card required.

32 GraphHub tools catalogued and risk-classified — across an index of 43,000+ MCP servers.