cve_id stringlengths 13 15 | title stringlengths 14 164 ⌀ | description stringlengths 62 2.11k | date_published stringlengths 24 24 | affected listlengths 1 5 | upstream_weaknesses listlengths 0 5 | upstream_severity_metrics listlengths 0 4 | vulnerability_type stringclasses 33
values | component stringclasses 7
values | root_cause listlengths 1 4 ⌀ | classification_confidence dict | mechanism_relation stringclasses 3
values | model_necessity stringclasses 3
values | sensitive_operation stringclasses 107
values | trust_boundary stringclasses 105
values | untrusted_entry stringclasses 111
values | evidence listlengths 2 11 | reasoning dict |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
CVE-2025-10193 | Neo4j Cypher MCP server is vulnerable to DNS rebinding attacks | DNS rebinding vulnerability in Neo4j Cypher MCP server allows malicious websites to bypass Same-Origin Policy protections and execute unauthorised tool invocations against locally running Neo4j MCP instances. The attack relies on the user being enticed to visit a malicious website and spend sufficient time there for DN... | 2025-09-11T14:05:30.592Z | [
{
"collectionURL": "https://pypi.python.org",
"cpes": null,
"defaultStatus": "unaffected",
"modules": null,
"packageName": "mcp-neo4j-cypher",
"packageURL": null,
"platforms": null,
"product": "neo4j-cypher MCP server",
"programFiles": null,
"programRoutines": null,
"repo... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-346",
"description": "CWE-346 Origin Validation Error",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": "NO",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "HIGH",
"attackRequirements": "PRESENT",
"attackVector": "NETWORK",
... | DNS Rebinding | server | [
"Protocol/Interaction Boundary Controls"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | Local Neo4j tool invocation | Web origin → local service | Malicious web page or DNS response | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: Neo4j Cypher MCP server provides tool/resource fun... |
CVE-2025-10619 | sequa-ai sequa-mcp OAuth Server Discovery node-oauth-client-provider.ts redirectToAuthorization os command injection | A vulnerability was detected in sequa-ai sequa-mcp up to 1.0.13. This affects the function redirectToAuthorization of the file src/helpers/node-oauth-client-provider.ts of the component OAuth Server Discovery. Performing manipulation results in os command injection. Remote exploitation of the attack is possible. The ex... | 2025-09-17T21:02:11.492Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": [
"OAuth Server Discovery"
],
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "sequa-mcp",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": ... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-78",
"description": "OS Command Injection",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
},
{
"description": null,
"descriptions": [
{
"cweId": "CWE-77"... | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": null,
"Recovery": null,
"Safety": null,
"attackComplexity": null,
"attackRequirements": null,
"attackVector": null,
"baseScore": 5.3,
"baseS... | Command Injection | client | [
"Input Validation"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "medium"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | Local URL launch / reported command execution | Remote authorization metadata → local launcher | OAuth discovery endpoint | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: The patch describes a local STDIO proxy connecting... |
CVE-2025-11285 | samanhappy MCPHub serverController.ts os command injection | A vulnerability was found in samanhappy MCPHub up to 0.9.10. Affected by this issue is some unknown functionality of the file src/controllers/serverController.ts. The manipulation of the argument command/args results in os command injection. The attack can be launched remotely. The exploit has been made public and coul... | 2025-10-05T06:02:06.272Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "MCPHub",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "samanhappy",
"versions": [
... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-78",
"description": "OS Command Injection",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
},
{
"description": null,
"descriptions": [
{
"cweId": "CWE-77"... | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": null,
"Recovery": null,
"Safety": null,
"attackComplexity": null,
"attackRequirements": null,
"attackVector": null,
"baseScore": 5.3,
"baseS... | Code Execution (Mechanism Unspecified) | client | null | {
"component_bucket": "medium",
"root_cause": "low",
"vulnerability_type": "medium"
} | null | not_required_for_documented_primitive | Launching a configured process | null | command/args in service configuration | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1",
"E2",
"E3",
"E4",
"E7",
"E9",
"E10",
"E11"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
... |
CVE-2025-11286 | samanhappy MCPHub MCPRouter Service serverController.ts server-side request forgery | A vulnerability was determined in samanhappy MCPHub up to 0.9.10. This affects an unknown part of the file src/controllers/serverController.ts of the component MCPRouter Service. This manipulation of the argument baseUrl causes server-side request forgery. The attack may be initiated remotely. The exploit has been publ... | 2025-10-05T06:32:05.838Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": [
"MCPRouter Service"
],
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "MCPHub",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "samanha... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-918",
"description": "Server-Side Request Forgery",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": null,
"Recovery": null,
"Safety": null,
"attackComplexity": null,
"attackRequirements": null,
"attackVector": null,
"baseScore": 5.1,
"baseS... | SSRF | client | [
"Input Validation"
] | {
"component_bucket": "medium",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | not_required_for_documented_primitive | Server-side network request | Client-supplied destination → server-side network access | baseUrl route parameter | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1",
"E2"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: The affected baseUrl configures the do... |
CVE-2025-11287 | samanhappy MCPHub sseService.ts handleSseConnectionfunction improper authentication | A vulnerability was identified in samanhappy MCPHub up to 0.9.10. This vulnerability affects the function handleSseConnectionfunction of the file src/services/sseService.ts. Such manipulation leads to improper authentication. The attack may be launched remotely. The exploit is publicly available and might be used. The ... | 2025-10-05T07:02:06.182Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "MCPHub",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "samanhappy",
"versions": [
... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-287",
"description": "Improper Authentication",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": null,
"Recovery": null,
"Safety": null,
"attackComplexity": null,
"attackRequirements": null,
"attackVector": null,
"baseScore": 6.9,
"baseS... | Authentication Bypass | server | [
"Authentication"
] | {
"component_bucket": "medium",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | Access to an MCP service | Unauthenticated caller → service session | SSE connection request | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1",
"E2"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: The named SSE connection handler creat... |
CVE-2025-13822 | Authentication bypass in MCPHub | MCPHub in versions below 0.11.0 is vulnerable to authentication bypass. Some endpoints are not protected by authentication middleware, allowing an unauthenticated attacker to perform actions in the name of other users and using their privileges. | 2026-04-14T10:23:49.910Z | [
{
"collectionURL": "https://www.npmjs.com/package/@samanhappy/mcphub",
"cpes": null,
"defaultStatus": "unaffected",
"modules": null,
"packageName": "@samanhappy/mcphub",
"packageURL": null,
"platforms": null,
"product": "MCPHub",
"programFiles": null,
"programRoutines": null,... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-639",
"description": "CWE-639 Authorization Bypass Through User-Controlled Key",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "ADJACENT",
... | Authentication Bypass | null | [
"Authentication"
] | {
"component_bucket": "low",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | not_required_for_documented_primitive | Operations under another user's privileges | Anonymous caller → user privileges | Request to an unprotected API | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1",
"E3",
"E5",
"E7",
"E8",
"E9",
"E10",
"E11"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
... |
CVE-2025-14660 | DecoCMS Mesh Workspace Domain api.ts createTool access control | A flaw has been found in DecoCMS Mesh up to 1.0.0-alpha.31. Affected by this vulnerability is the function createTool of the file packages/sdk/src/mcp/teams/api.ts of the component Workspace Domain Handler. This manipulation of the argument domain causes improper access controls. The attack can be initiated remotely. T... | 2025-12-14T12:32:08.752Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": [
"Workspace Domain Handler"
],
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "Mesh",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "De... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-284",
"description": "Improper Access Controls",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
},
{
"description": null,
"descriptions": [
{
"cweId": "CW... | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": null,
"Recovery": null,
"Safety": null,
"attackComplexity": null,
"attackRequirements": null,
"attackVector": null,
"baseScore": 6.3,
"baseS... | Authorization Bypass | server | [
"Authorization"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | Joining or accessing a team workspace | Caller → another team's privileges | domain parameter of createTool | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1",
"E2"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: The pinned fix and teams/api.ts identi... |
CVE-2025-15061 | Framelink Figma MCP Server fetchWithRetry Command Injection Remote Code Execution Vulnerability | Framelink Figma MCP Server fetchWithRetry Command Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Framelink Figma MCP Server. Authentication is not required to exploit this vulnerability.
The specific flaw exists within th... | 2026-01-23T03:20:19.025Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": "unknown",
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "Figma MCP Server",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "Framelink",
"vers... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-78",
"description": "CWE-78: Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": {
"baseScore": 9.8,
"baseSeverity": "CRITICAL",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"cvssV3_1": null,
"cvssV4_0": null,
"format": "CVSS",
"other": null,
... | Command Injection | server | [
"Input Validation"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | not_required_for_documented_primitive | System commands under the service account | Input data → shell syntax | Input to fetchWithRetry | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: Figma MCP Server provides tool implementation for ... |
CVE-2025-15063 | Ollama MCP Server execAsync Command Injection Remote Code Execution Vulnerability | Ollama MCP Server execAsync Command Injection Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Ollama MCP Server. Authentication is not required to exploit this vulnerability.
The specific flaw exists within the implementation of the... | 2026-01-23T02:47:32.505Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": "unknown",
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "Ollama MCP Server",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "Ollama MCP Server",
... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-78",
"description": "CWE-78: Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": {
"baseScore": 9.8,
"baseSeverity": "CRITICAL",
"vectorString": "CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H",
"version": "3.0"
},
"cvssV3_1": null,
"cvssV4_0": null,
"format": "CVSS",
"other": null,
... | Command Injection | server | [
"Input Validation"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | not_required_for_documented_primitive | System commands under the service account | Input data → shell syntax | Input to execAsync | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: Ollama MCP Server provides server-side command exe... |
CVE-2025-20381 | SPL commands allowlist controls bypass in Splunk MCP Server app through "run_splunk_query" MCP tool | In Splunk MCP Server app versions below 0.2.4, a user with access to the "run_splunk_query" Model Context Protocol (MCP) tool could bypass the SPL command allowlist controls in MCP by embedding SPL commands as sub-searches, leading to unauthorized actions beyond the intended MCP restrictions. | 2025-12-03T17:00:25.945Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "Splunk MCP Server",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "Splunk",
"versions": ... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-863",
"description": "The software performs an authorization check when an actor attempts to access a resource or perform an action, but it does not correctly perform the check. This allows attackers to bypass intended access res... | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": null,
"attackVector": null,
"availabilityImpact": null,
"baseScore": 5.4,
"baseSeverity": "MEDIUM",
"confidentialityImpact": null,
"exploitCodeMaturity": null,
... | Authorization Bypass | server | [
"Authorization"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | Executing query operations outside an allowlist | Authorized tool → unauthorized sub-operation | SPL query and subsearches | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: Splunk MCP Server provides run_splunk_query tool w... |
CVE-2025-34072 | Anthropic Slack MCP Server Data Exfiltration via Link Unfurling | A data exfiltration vulnerability exists in Anthropic’s deprecated Slack Model Context Protocol (MCP) Server via automatic link unfurling. When an AI agent using the Slack MCP Server processes untrusted data, it can be manipulated to generate messages containing attacker-crafted hyperlinks embedding sensitive data. Sla... | 2025-07-02T13:46:31.554Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": "unaffected",
"modules": [
"Anthropic Slack MCP Server",
"Message content (post/reply body)"
],
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "Slack MCP Server",
"programFiles": null,... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-200",
"description": "CWE-200 Exposure of Sensitive Information to an Unauthorized Actor",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
},
{
"description": null,
"des... | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
... | Sensitive Data Exposure | server | [
"Sensitive Data Handling"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | agent_or_tool_output_composition | required_for_documented_agent_chain | Sending Slack messages containing secret links / automatic preview requests to external destinations | External data → agent instructions → outbound disclosure | Untrusted content read by an agent | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: Anthropic Slack MCP Server posts messages to Slack... |
CVE-2025-35028 | HexStrike AI MCP Server Command Injection | By providing a command-line argument starting with a semi-colon ; to an API endpoint created by the EnhancedCommandExecutor class of the HexStrike AI MCP server, the resultant composed command is executed directly in the context of the MCP server’s normal privilege; typically, this is root. There is no attempt to sanit... | 2025-11-30T21:27:56.057Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": "unaffected",
"modules": [
"EnhancedCommandExecutor"
],
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "HexStrike AI",
"programFiles": null,
"programRoutines": null,
"repo": null,
... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-78",
"description": "CWE-78 Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "NONE",
"baseScore": 9.1,
"baseSeverity": "CRITICAL",
"confidentialityImpact": "HIGH",
"exploitCodeMaturity... | Command Injection | server | [
"Input Validation"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | not_required_for_documented_primitive | Shell execution in the service process | Argument data → command syntax | API command-line arguments | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: EnhancedCommandExecutor in HexStrike AI MCP server... |
CVE-2025-4143 | Missing validation of redirect_uri on authorize endpoint | The OAuth implementation in workers-oauth-provider that is part of MCP framework https://github.com/cloudflare/workers-mcp , did not correctly validate that redirect_uri was on the allowed list of redirect URIs for the given client registration.
Fixed in: https://github.com/cloudflare/workers-oauth-provider/pull/26 ... | 2025-05-01T00:19:52.737Z | [
{
"collectionURL": "https://github.com/cloudflare",
"cpes": null,
"defaultStatus": "unaffected",
"modules": null,
"packageName": "workers-oauth-provider",
"packageURL": null,
"platforms": null,
"product": null,
"programFiles": null,
"programRoutines": null,
"repo": "https... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-601",
"description": "CWE-601 URL Redirection to Untrusted Site ('Open Redirect')",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "HIGH",
"attackRequirements": "PRESENT",
"attackVector": "NETWORK... | Session or Peer Binding Failure | sdk | [
"Session or Peer Binding"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | Sending authorization credentials | Registered client → callback recipient | OAuth authorization request / malicious callback | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: workers-oauth-provider is MCP-framework OAuth SDK ... |
CVE-2025-4144 | PKCE bypass via downgrade attack | PKCE was implemented in the OAuth implementation in workers-oauth-provider that is part of MCP framework https://github.com/cloudflare/workers-mcp . However, it was found that an attacker could cause the check to be skipped.
Fixed in:
https://github.com/cloudflare/workers-oauth-provider/pull/27 https://github.c... | 2025-05-01T00:50:27.543Z | [
{
"collectionURL": "https://github.com/cloudflare",
"cpes": null,
"defaultStatus": "unaffected",
"modules": null,
"packageName": "workers-oauth-provider",
"packageURL": null,
"platforms": null,
"product": null,
"programFiles": null,
"programRoutines": null,
"repo": "https... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-287",
"description": "CWE-287 Improper Authentication",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": "NOT_DEFINED",
"Recovery": "NOT_DEFINED",
"Safety": "NOT_DEFINED",
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
... | Session or Peer Binding Failure | sdk | [
"Session or Peer Binding"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | Bypassing PKCE validation | Authorization initiator → code redeemer | OAuth authorization exchange parameters | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: workers-oauth-provider is MCP-framework OAuth SDK ... |
CVE-2025-47274 | ToolHive stores secrets in the state store with no encryption | ToolHive is a utility designed to simplify the deployment and management of Model Context Protocol (MCP) servers. Due to the ordering of code used to start an MCP server container, versions of ToolHive prior to 0.0.33 inadvertently store secrets in the run config files which are used to restart stopped containers. This... | 2025-05-12T14:57:46.781Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "toolhive",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "stacklok",
"versions": [
... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-311",
"description": "CWE-311: Missing Encryption of Sensitive Data",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": null,
"Recovery": null,
"Safety": null,
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "LOCAL",
"baseScore": 2.4,
... | Sensitive Data Exposure | client | [
"Sensitive Data Handling"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | not_required_for_documented_primitive | Reading persisted secrets | Secret storage → ordinary user-directory file | Reading a local runconfig file | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: ToolHive is orchestrator/manager of MCP server dep... |
CVE-2025-47777 | 5ire Client Vulnerable to Cross-Site Scripting (XSS) and Remote Code Execution (RCE) | 5ire is a cross-platform desktop artificial intelligence assistant and model context protocol client. Versions prior to 0.11.1 are vulnerable to stored cross-site scripting in chatbot responses due to insufficient sanitization. This, in turn, can lead to Remote Code Execution (RCE) via unsafe Electron protocol handling... | 2025-05-14T15:23:28.565Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "5ire",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "nanbingxyz",
"versions": [
{... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-20",
"description": "CWE-20: Improper Input Validation",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
},
{
"description": null,
"descriptions": [
{
"cwe... | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 9.7,
"baseSeverity": "CRITICAL",
"confidentialityImpact": "HIGH",
"exploitCodeMaturity... | XSS | user-interface | [
"Input Validation"
] | {
"component_bucket": "medium",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | not_required_for_documented_primitive | Script execution and Electron capabilities | Untrusted HTML → privileged desktop API | Chat response / pasted external content | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "The affected chatbot-response HTML renderer is application presentation c... |
CVE-2025-49596 | MCP Inspector proxy server lacks authentication between the Inspector client and proxy | The MCP inspector is a developer tool for testing and debugging MCP servers. Versions of MCP Inspector below 0.14.1 are vulnerable to remote code execution due to lack of authentication between the Inspector client and proxy, allowing unauthenticated requests to launch MCP commands over stdio. Users should immediately ... | 2025-06-13T20:11:40.453Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "inspector",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "modelcontextprotocol",
"versi... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-306",
"description": "CWE-306: Missing Authentication for Critical Function",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": null,
"Recovery": null,
"Safety": null,
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 9.4,
... | Authentication Bypass | client | [
"Authentication"
] | {
"component_bucket": "medium",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | Launching a stdio MCP command | Unauthenticated client → local process capabilities | Inspector proxy request | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: The affected Inspector proxy launches MCP commands... |
CVE-2025-5071 | AI Engine 2.8.0 - 2.8.3 - Authenticated (Subscriber+) Insufficient Authorization to Privilege Escalation via MCP | The AI Engine plugin for WordPress is vulnerable to unauthorized modification of data and loss of data due to a missing capability check on the 'Meow_MWAI_Labs_MCP::can_access_mcp' function in versions 2.8.0 to 2.8.3. This makes it possible for authenticated attackers, with subscriber-level access and above, to have fu... | 2025-06-19T09:23:47.875Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": "unaffected",
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "AI Engine",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "tigroumeow",
"version... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-863",
"description": "CWE-863 Incorrect Authorization",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": null,
"attackVector": null,
"availabilityImpact": null,
"baseScore": 8.8,
"baseSeverity": "HIGH",
"confidentialityImpact": null,
"exploitCodeMaturity": null,
... | Authorization Bypass | server | [
"Authorization"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | WordPress user/content modification | Subscriber identity → administrative capabilities | MCP request from a low-privilege user | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: AI Engine MCP integration in WordPress plugin prov... |
CVE-2025-52573 | Command Injection in MCP Server ios-simulator-mcp | iOS Simulator MCP Server (ios-simulator-mcp) is a Model Context Protocol (MCP) server for interacting with iOS simulators. Versions prior to 1.3.3 are written in a way that is vulnerable to command injection vulnerability attacks as part of some of its MCP Server tool definition and implementation. The MCP Server expos... | 2025-06-26T14:08:56.100Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "ios-simulator-mcp",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "joshuayoes",
"version... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-78",
"description": "CWE-78: Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": "HIGH",
"attackVector": "LOCAL",
"availabilityImpact": "HIGH",
"baseScore": 6,
"baseSeverity": "MEDIUM",
"confidentialityImpact": "NONE",
"exploitCodeMaturity": nu... | Command Injection | server | [
"Input Validation"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | optional_delivery_route | idb execution through a shell | Tool arguments → shell syntax | ui_tap tool arguments | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: ios-simulator-mcp MCP server provides iOS simulato... |
CVE-2025-5273 | null | Versions of the package mcp-markdownify-server before 1.0.0 are vulnerable to Files or Directories Accessible to External Parties via the get-markdown-file tool. An attacker can craft a prompt that, once accessed by the MCP host, will allow it to read arbitrary files from the host running the server. | 2025-05-29T05:00:01.272Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "mcp-markdownify-server",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "n/a",
"versions"... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-552",
"description": "Files or Directories Accessible to External Parties",
"lang": "en",
"type": null
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "NONE",
"baseScore": 6.5,
"baseSeverity": "MEDIUM",
"confidentialityImpact": "HIGH",
"exploitCodeMaturity":... | Path Traversal | server | [
"Input Validation",
"Authorization"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | optional_delivery_route | Reading files outside the directory | Allowed file scope → arbitrary files | get-markdown-file path | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: mcp-markdownify-server provides file retrieval too... |
CVE-2025-5276 | null | Versions of the package mcp-markdownify-server before 1.0.0 are vulnerable to Server-Side Request Forgery (SSRF) via the Markdownify.get() function. An attacker can craft a prompt that, once accessed by the MCP host, can invoke the webpage-to-markdown, bing-search-to-markdown, and youtube-to-markdown tools to issue req... | 2025-05-29T05:00:05.001Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "mcp-markdownify-server",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "n/a",
"versions"... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-918",
"description": "Server-Side Request Forgery (SSRF)",
"lang": "en",
"type": null
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "NONE",
"baseScore": 7.4,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"exploitCodeMaturity": "... | SSRF | server | [
"Input Validation"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | optional_delivery_route | Server-side fetch and response reading | External URL → internal network resources | URL supplied to a web-conversion tool | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: mcp-markdownify-server provides Markdownify tools ... |
CVE-2025-5277 | null | aws-mcp-server MCP server is vulnerable to command injection. An attacker can craft a prompt that once accessed by the MCP client will run arbitrary commands on the host system. | 2025-05-28T13:14:26.863Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "aws-mcp-server",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "alexei-led",
"versions":... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-78",
"description": "OS Command Injection",
"lang": "en",
"type": null
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": "LOW",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 9.6,
"baseSeverity": "CRITICAL",
"confidentialityImpact": "HIGH",
"exploitCodeMaturity... | Command Injection | server | [
"Input Validation"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | optional_delivery_route | Local shell execution | Restricted AWS / pipeline commands → shell capability | Tool command / pipeline input, potentially guided by prompts | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1",
"E2"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: aws-mcp-server executes AWS CLI comman... |
CVE-2025-53098 | Roo Code Vulnerable to Potential Remote Code Execution via Model Context Protocol | Roo Code is an AI-powered autonomous coding agent. The project-specific MCP configuration for the Roo Code agent is stored in the `.roo/mcp.json` file within the VS Code workspace. Because the MCP configuration format allows for execution of arbitrary commands, prior to version 3.20.3, it would have been possible for a... | 2025-06-27T21:43:35.031Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "Roo-Code",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "RooCodeInc",
"versions": [
... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-77",
"description": "CWE-77: Improper Neutralization of Special Elements used in a Command ('Command Injection')",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": "HIGH",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 8.1,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"exploitCodeMaturity": ... | Configuration Trust or Approval Bypass | client | [
"User Approval and Configuration Trust"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | agent_or_tool_output_composition | required_for_documented_agent_chain | Writing .roo/mcp.json and launching a command | Ordinary file write → executable-configuration approval | Attacker prompt / indirect prompt | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: Roo Code is VS Code extension/client agent managin... |
CVE-2025-53100 | RestDB's Codehooks.io MCP Server Vulnerable to Command Injection | RestDB's Codehooks.io MCP Server is an MCP server on the Codehooks.io platform. Prior to version 0.2.2, the MCP server is written in a way that is vulnerable to command injection attacks as part of some of its MCP Server tools definition and implementation. This could result in a user initiated remote command injection... | 2025-07-01T17:49:54.733Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "codehooks-mcp-server",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "RestDB",
"versions... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-78",
"description": "CWE-78: Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": null,
"Recovery": null,
"Safety": null,
"attackComplexity": "LOW",
"attackRequirements": "NONE",
"attackVector": "NETWORK",
"baseScore": 8.6,
... | Command Injection | server | [
"Input Validation"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | not_required_for_documented_primitive | child_process.exec | Input data → shell syntax | Codehooks tool input | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: Codehooks.io MCP Server provides tools invoked by ... |
CVE-2025-53107 | @cyanheads/git-mcp-server vulnerable to command injection in several tools | @cyanheads/git-mcp-server is an MCP server designed to interact with Git repositories. Prior to version 2.1.5, there is a command injection vulnerability caused by the unsanitized use of input parameters within a call to child_process.exec, enabling an attacker to inject arbitrary system commands. Successful exploitati... | 2025-07-01T17:55:30.726Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "git-mcp-server",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "cyanheads",
"versions": ... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-77",
"description": "CWE-77: Improper Neutralization of Special Elements used in a Command ('Command Injection')",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": {
"attackComplexity": "HIGH",
"attackVector": "NETWORK",
"availabilityImpact": "HIGH",
"baseScore": 7.5,
"baseSeverity": "HIGH",
"confidentialityImpact": "HIGH",
"exploitCodeMaturity": ... | Command Injection | server | [
"Input Validation"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_reuse | optional_delivery_route | Git command execution through a shell | Tool data → shell syntax | Git tool arguments / untrusted Git logs | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: git-mcp-server provides git command tools called b... |
CVE-2025-53109 | Model Context Protocol Servers Vulnerable to Path Validation Bypass via Prefix Matching and Symlink Handling | Model Context Protocol Servers is a collection of reference implementations for the model context protocol (MCP). Versions of Filesystem prior to 0.6.4 or 2025.7.01 could allow access to unintended files via symlinks within allowed directories. Users are advised to upgrade to 0.6.4 or 2025.7.01 resolve. | 2025-07-02T14:30:57.647Z | [
{
"collectionURL": null,
"cpes": null,
"defaultStatus": null,
"modules": null,
"packageName": null,
"packageURL": null,
"platforms": null,
"product": "servers",
"programFiles": null,
"programRoutines": null,
"repo": null,
"vendor": "modelcontextprotocol",
"version... | [
{
"description": null,
"descriptions": [
{
"cweId": "CWE-59",
"description": "CWE-59: Improper Link Resolution Before File Access ('Link Following')",
"lang": "en",
"type": "CWE"
}
],
"source": null,
"type": null
}
] | [
{
"cvssMetricV31": null,
"cvssV2_0": null,
"cvssV3_0": null,
"cvssV3_1": null,
"cvssV4_0": {
"Automatable": null,
"Recovery": null,
"Safety": null,
"attackComplexity": "LOW",
"attackRequirements": "PRESENT",
"attackVector": "NETWORK",
"baseScore": 7.3,
... | Path Traversal | server | [
"Input Validation",
"Authorization"
] | {
"component_bucket": "high",
"root_cause": "high",
"vulnerability_type": "high"
} | traditional_weakness_at_mcp_integration_boundary | not_required_for_documented_primitive | Out-of-scope file access | Allowed directory → actual target path | File-tool path / symbolic link | [
{
"id": "E0",
"kind": "cve_record",
"related_urls": [],
"summary": "Source for basic fields. The exported values reflect a saved catalog snapshot; the linked online record may change.",
"supports": [
"title",
"description",
"affected",
"date_published",
"upstream_we... | {
"component": {
"basis": "analyst_classification",
"evidence_ids": [
"E1"
],
"limitation": "Summary of the existing classification review; public evidence links do not constitute independent reproduction.",
"reason": "v2.6 individual check: Filesystem server provides file access tools to MC... |
End of preview. Expand in Data Studio
MCP Vulnerability Dataset
791 CVE records related to Model Context Protocol (MCP), with 18 fields per
record and English evidence-based annotations. Data release: full-791-v1;
taxonomy: mcp-cve-taxonomy-v2.6; mechanism definitions: v1.
Explore and use
- Browse all 791 records in the native Dataset Viewer.
- Reading guide and field reference: a beginner-friendly example, all field definitions, evidence semantics and Python usage.
- Taxonomy: vulnerability types, component roles, root causes and classification boundaries.
- Download JSONL, JSON, or Parquet. These contain the same 791 records; Parquet supplies the Viewer and pads absent optional object keys with null.
The single full split contains all 791 records. No custom dataset loader is required.
Scope and limitations
This is a frozen research cohort, not an exhaustive census or a live security feed. MCP relevance does not imply a defect in the MCP specification or dependence on a language model. Multiple CVEs may describe overlapping mechanisms.
Annotations derive from AI-assisted static review of saved CVE records, advisories and available code or patches, without independent human validation or a claim of successful exploit reproduction. Upstream CWE/CVSS assessments remain separate from project judgments. Evidence depth varies; unsupported conclusions remain unknown. Linked sources can change, so recheck current advisories for deployment decisions. Full original-language audits, source snapshots and change records remain local; this public package excludes full third-party reports and code.
Data coverage and version history
Data snapshot / collection cutoff: 2026-10-24 (UTC). Included CVEs have official first-publication dates from 2025-05-01 through 2026-10-23. The latest is CVE-2026-115845, published at 2026-10-23 19:04:44.790 UTC. The cutoff describes the collection snapshot, not guaranteed coverage of every CVE published through the end of October 24. It is separate from publication or documentation-update dates.
Mechanism annotations were completed on 2026-11-03 using saved evidence; that review did not refresh the entire CVE cohort. The English full-cohort release was published on 2026-11-07. Later catalog updates are not included unless a new data release explicitly changes the snapshot date.
| Version / update | Published (UTC date) | Data snapshot (UTC date) | Records | Change |
|---|---|---|---|---|
| Repository and split naming update | 2026-11-07 | 2026-10-24 | 791 | Repository renamed to mcp-vulnerability-dataset; split renamed to full; record files unchanged. |
| Documentation update | 2026-11-07 | 2026-10-24 | 791 | Consolidated field reference into the guide; removed duplicate public documents. |
| full-791-v1 | 2026-11-07 | 2026-10-24 | 791 | Full English cohort; supplemental evidence bindings and reversible metric-shape normalization. |
| Accepted English preview | 2026-11-07 | 2026-10-24 | 10 | Initial presentation and beginner's guide. |
Future entries will record both the data snapshot date and release date, the row
count, and material data or annotation changes. Documentation-only updates do not
advance the data cutoff. Historical revisions retain their original split names;
use full for the current release.
The underlying catalog is pinned to project source revision
554b915689d27a1e8810bfdf0aab1d77358557c6. The full release passed 791-record,
18-field consistency checks and verification of 1,187 saved evidence hashes.
Historical public versions remain available through repository history. No
automatic Hugging Face synchronization is enabled.
License and attribution
See LICENSE, NOTICE and upstream terms. CC BY 4.0 applies only to original contributions to the extent rights exist; CVE material retains MITRE's terms, and third-party rights remain unchanged. Attribution: MCP Vulnerability Dataset contributors (published by XuViewer).
- Downloads last month
- 127