// SCAN REPORT

Musubix

126 tools. 47 can modify or destroy data without limits.

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

Last updated: 28 Jun 2026

47 can modify or destroy data

79 read-only

126 tools total

Community server · catalogue entry verified 28/06/2026

How to control Musubix ↓

TOOL MAP

What Musubix exposes to your agents

Read (79) Write / Execute (46) Destructive / Financial (1)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous Musubix tools

Destructive · High knowledge_delete_entity Deletion of knowledge graph entities cannot be undone and permanently removes data. Execute · Medium agent_cancel This tool cancels running agent tasks and interrupts subtasks. Execute · Medium assistant_axis_session_start This tool executes an operation that initiates a monitoring session, which has side effects (session creation, state initialization). Execute · Medium codegraph_index While the tool does not delete data or move money, it executes a non-trivial computational operation that analyzes and processes code.

47 of Musubix's 126 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Musubix

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

Deny destructive operations

{
  "knowledge_delete_entity": {
    "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

{
  "assistant_axis_session_end": {
    "limits": [
      {
        "counter": "assistant_axis_session_end_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "agent_analyze": {
    "limits": [
      {
        "counter": "agent_analyze_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 Musubix — 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 MUSUBIX →

Instant setup, no code required.

FULL CATALOGUE

All 126 Musubix tools

DESTRUCTIVE 1 tools

Destructive knowledge_delete_entity Delete an entity from the knowledge graph

EXECUTE 25 tools

Execute agent_cancel 実行中のエージェントタスクをキャンセルします。まだ開始されていないサブタスクはスキップされ、実行中のタスクは可能な限り中断されます。 Execute agent_dispatch サブエージェントをディスパッチして複雑なタスクを実行します。タスクの複雑度を分析し、必要に応じてサブエージェントに分解して並列実行します。知識ストアとの統合により、実行中のコンテキスト共有を行います。使用例 Execute assistant_axis_session_start Start a new persona monitoring session. Returns session ID and initial state. Execute codegraph_index Index a code repository or directory to build a code graph. This extracts entities (classes, functions, etc.) Execute deep_research_start Start iterative deep research on a technical topic using search-read-reason cycle Execute delegation_retry Retry a failed delegation with optional feedback. Implements exponential backoff. Execute ears_to_smt Convert an EARS (Easy Approach to Requirements Syntax) requirement to SMT-LIB2 formula for formal verification Execute expert_architect Delegate architecture-related tasks to the Architect expert. Use for system design, component structure, desig Execute expert_delegate Delegate a task to a specialized AI expert. The system automatically selects the appropriate expert based on t Execute expert_ears Delegate requirements analysis to the EARS Analyst expert. Use for converting natural language requirements to Execute expert_formal Delegate formal verification to the Formal Verifier expert. Use for generating Hoare logic conditions, Z3/SMT Execute expert_security Delegate security analysis to the Security Analyst expert. Use for vulnerability detection, security code revi Execute sdd_formal_verify Perform formal verification on EARS requirements using Z3 theorem prover. Execute security_generate_fix Generate a fix for a specific vulnerability. Execute security_scan Run a comprehensive security scan on the target path. Detects vulnerabilities, hardcoded secrets, tainted data Execute security_taint_analysis Run taint analysis to trace data flow from untrusted sources to sensitive sinks. Execute skill_execute 指定したスキルを実行します。スキルのパラメータを指定して実行し、結果を返します。 Execute symbolic_filter_code Filter LLM-generated code through symbolic validation pipeline (hallucination detection, constitution complian Execute synthesis_from_examples Synthesize a program from input/output examples using Programming by Example (PBE) Execute verify_postcondition Verify a Hoare triple {P} C {Q} using Z3 SMT solver. Checks if the postcondition Q holds after executing C whe Execute verify_precondition Verify a precondition expression using Z3 SMT solver. Checks if the precondition is satisfiable or unsatisfiab Execute watch_run_now Run specified tasks immediately on given files without waiting for changes. Execute watch_start Start file watching and auto-validation. Monitors files for changes and runs lint/test/security automatically. Execute workflow_gate フェーズのクオリティゲートを実行します。フェーズごとに定義された品質チェックを実行し、結果を返します。 Execute workflow_transition ワークフローを次のフェーズに遷移します。現在のフェーズが承認された後にのみ遷移可能です。ユーザーからの承認キーワード（承認, OK, LGTM, 進める）が必要です。 ⚠️ Phase 2 (設計) から P

WRITE 21 tools

Write assistant_axis_session_end End a monitoring session and get final metrics summary. Write decision_accept Accept a proposed ADR Write decision_create Create a new Architecture Decision Record (ADR) Write decision_generate_index Generate or update the ADR index file Write knowledge_add_relation Add a relation between two entities Write knowledge_put_entity Add or update an entity in the knowledge graph Write pattern_consolidate Consolidate and optimize the pattern library. Part of the Sleep phase of Wake-Sleep learning. Write pattern_import_kg Import learned patterns into the ontology knowledge graph for semantic reasoning. Write provider_select Select and configure the LLM provider for expert delegation. Currently supports VS Code Language Model API. Write sdd_create_design Create a C4 model design document Write sdd_create_requirements Create a new EARS-format requirements document. Returns clarifying questions if context is incomplete. Write sdd_create_tasks Create implementation tasks from design Write sdd_update_knowledge Update or add nodes to the MUSUBIX knowledge graph Write skill_register 新しいスキルを登録します。カスタムスキルを登録して利用可能にします。 Write spaces_activate スペースをアクティベートし、関連ファイルとコンテキストを読み込みます。 Write spaces_create 新しいCopilot Spaceを作成します。要件、設計、タスク、ファイルを関連付けて、コンテキストを管理できます。 Write team_add_knowledge チームのナレッジベースに新しいエントリを追加します。決定、教訓、ベストプラクティス、警告などを共有できます。 Write team_share_pattern チームとコードパターンを共有します。Gitブランチ経由で共有され、チームメンバーが採用できます。 Write team_sync チームのパターンとナレッジをリモートと同期します。 Write trace_add_link Add a traceability link between two artifacts (requirements, designs, code, tests). Write workflow_create 新しいワークフローを作成し、Phase 1（要件定義）を開始します。 SDDワークフローは以下の5フェーズで構成されます: 1. 要件定義 (EARS形式) 2. 設計 (C4モデル) 3. タスク分解 (必須 -

READ 79 tools

Read agent_analyze タスクの複雑度を分析し、サブエージェント分解の推奨を提供します。 5つの因子（スコープ、依存関係、ファイル数、テストカバレッジ、不確実性）で分析し、分解が必要かどうかを判定します。 Read agent_status ディスパッチされたエージェントの状態を取得します。実行中のサブエージェントの進捗、完了状況、エラー情報を提供します。 Read assistant_axis_analyze Analyze a message for persona drift indicators. Returns drift score, detected triggers, domain classification, Read assistant_axis_config Get current Assistant Axis configuration including thresholds and monitoring levels. Read assistant_axis_get_reinforcement Get identity reinforcement prompt for the current session state. Used when drift is detected. Read assistant_axis_phase_check Check monitoring level and frequency for a specific workflow phase. Read assistant_axis_session_status Get current status of a monitoring session including drift score, trend, and intervention count. Read check_circular Check for circular dependencies in relationships. Detects cycles in dependency graphs that could cause infini Read codegraph_find_callees Find all functions/methods called by a specific entity Read codegraph_find_callers Find all entities that call a specific function or method Read codegraph_find_dependencies Find all dependencies (imports, extends, implements, calls) of a code entity Read codegraph_global_search Perform a semantic search across the entire codebase using GraphRAG. Best for high-level questions about archi Read codegraph_local_search Search for related code around a specific entity. Best for understanding how a particular piece of code fits i Read codegraph_query Query the code graph for entities matching specific criteria (type, name pattern, file path, etc.) Read codegraph_stats Get statistics about the indexed code graph (entity counts, relation counts, etc.) Read codeql_aggregate Aggregate multiple CodeQL scan results into a unified report with statistics. Read codeql_cwe_lookup Look up CWE information including Japanese explanations and recommended fixes. Read codeql_fix_suggestions Get fix suggestions for security findings based on CWE information. Read codeql_list_cwes List all known CWE entries, optionally filtered by category or severity. Read codeql_parse_sarif Parse a CodeQL SARIF 2.1.0 report file and extract security findings with CWE mappings. Read codeql_summary Get a summary of security findings from parsed SARIF data. Read consistency_validate Validate knowledge graph consistency using OWL constraints. Checks for: - Duplicate triples (exact and semant Read decision_deprecate Deprecate an ADR, optionally marking it as superseded Read decision_find_by_requirement Find ADRs related to a requirement Read decision_get Get a specific ADR by ID Read decision_list List all ADRs with optional filters Read decision_search Search ADRs by keyword Read deep_research_report Get final report from completed deep research session Read deep_research_status Get current status and progress of a deep research session Read expert_ontology Delegate ontology reasoning to the Ontology Reasoner expert. Use for knowledge graph queries, inference, and c Read expert_plan Delegate plan review to the Plan Reviewer expert. Use for validating designs, requirements, and implementation Read expert_review Delegate code review to the Code Reviewer expert. Use for code quality analysis, best practices review, and re Read knowledge_get_entity Get an entity from the knowledge graph by ID Read knowledge_query Query entities from the knowledge graph with filters Read knowledge_traverse Traverse the knowledge graph from a starting entity Read pattern_export_kg Export the pattern knowledge graph as Turtle RDF. Read pattern_learn Learn patterns from observed code. Part of the Wake phase of Wake-Sleep learning. Read pattern_query_relations Query relationships between patterns in the knowledge graph. Read pattern_search Search for patterns by various criteria. Read pattern_stats Get statistics about the pattern learning system. Read policy_check_file Check a specific file against relevant policies Read policy_get Get details of a specific policy Read policy_list List all registered policies Read policy_validate Validate a project against all registered policies Read sdd_query_knowledge Query the MUSUBIX knowledge graph for patterns, requirements, designs, and other artifacts Read sdd_validate_constitution Validate against the 9 Constitutional Articles Read sdd_validate_design Validate design against requirements traceability Read sdd_validate_requirements Validate requirements against EARS patterns and constitution Read sdd_validate_traceability Validate requirement-design-task traceability Read security_audit_deps Audit project dependencies for known vulnerabilities. Read security_detect_secrets Detect hardcoded secrets like API keys, passwords, and tokens. Read security_generate_report Generate a security report from scan results. Read security_quick_scan Quick vulnerability scan without taint analysis, secret detection, or dependency audit. Read skill_info スキルの詳細情報を取得します。パラメータ定義、タグ、使用例などの詳細を表示します。 Read skill_list 利用可能なスキルの一覧を取得します。スキルタイプやタグでフィルタリングできます。利用可能なスキルタイプ: - file-operation: ファイル操作 - code-analysis: コード分析 - c Read skill_validate スキル定義を検証します。メタデータの完全性、パラメータ定義の妥当性をチェックします。 Read spaces_list 全てのスペースを一覧表示します。タイプやキーワードでフィルタリングできます。 Read spaces_status 現在のスペースのステータス（アクティブスペース、統計情報）を表示します。 Read spaces_suggest クエリに基づいてコンテキストに追加すべきファイルを提案します。 Read symbolic_check_constitution Check code compliance against the 9 constitutional articles Read symbolic_detect_hallucinations Detect hallucinated references (non-existent APIs, modules, symbols) in generated code Read symbolic_estimate_confidence Estimate confidence score (0.0-1.0) for generated code with breakdown by syntactic, semantic, factual, and con Read symbolic_pipeline_info Get information about the symbolic processing pipeline and its configuration Read synthesis_analyze_examples Analyze the quality of input/output examples for synthesis Read synthesis_get_stats Get statistics about synthesis operations and pattern library Read synthesis_learn_patterns Learn reusable patterns from code snippets Read synthesis_query_patterns Query the pattern library for reusable patterns Read team_list_patterns チームで共有されているパターンの一覧を表示します。タイプやタグでフィルタリングできます。 Read team_query_knowledge チームのナレッジベースを検索します。タイプ、カテゴリ、タグでフィルタリングできます。 Read team_status チーム共有の現在のステータス（パターン数、ナレッジ数、Git状態）を表示します。 Read trace_impact Analyze the impact of changing an artifact. Returns all affected artifacts with impact scores. Read trace_query Query the traceability database to find artifacts and their relationships. Read trigger_detect Detect which expert should handle a given message. Returns the matched expert and confidence score. Read validate_triple Validate a single triple before adding to knowledge graph. Quick validation for individual data entries. Retu Read watch_report Get accumulated issues report from watch session. Read watch_status Get the status of active watch sessions. Read watch_stop Stop the active file watching session. Read workflow_review 現在のフェーズのセルフレビューを実行し、結果を表示します。レビュー完了後、ユーザーに承認/修正を求めます。 Read workflow_status ワークフローの現在の状態を取得します。各フェーズの状態、進捗率、アーティファクト情報を含みます。

EXPLORE

Related servers

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

FAQ

Questions about Musubix

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

Yes. The Musubix server exposes 1 destructive tools including knowledge_delete_entity. 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 Musubix? +

The Musubix server has 21 write tools including assistant_axis_session_end, decision_accept, decision_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 Musubix.

How many tools does the Musubix MCP server expose? +

126 tools across 4 categories: Destructive, Execute, Read, Write. 79 are read-only. 47 can modify, create, or delete data.

How do I enforce a policy on Musubix? +

Register the Musubix 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 Musubix tool call.

Deterministic rules across all 126 Musubix tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON MUSUBIX →

Instant setup, no code required.

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

// WHERE THIS COMES FROM

These policies come from Musubix'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 →