Security audit

Mcp Review

Security checks for vulnerabilities and agentic risk

Overview

This is a read-only MCP tool-design review skill with some wording-quality caveats but no hidden execution, persistence, or data-exfiltration behavior.

Install this if you want a Chinese-language checklist-based reviewer for MCP tool definitions. Be aware that its guidance favors adding trigger phrases to tool descriptions, so users reviewing high-impact write tools should still ensure descriptions include clear preconditions and confirmation expectations.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The guide explicitly encourages embedding natural-language trigger phrases in tool descriptions, but it does not pair that advice with scope boundaries, disambiguation rules, or negative examples. In an MCP setting, descriptions act as routing signals for the LLM, so broad triggers can cause the model to invoke tools on casual or only partially related user utterances, increasing the risk of unintended actions or incorrect tool selection.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The example phrases such as '帮我领券' and '一键领券' are common conversational expressions that may appear in broader discussion, making them poor standalone activation guidance. Because the document presents them as positive examples without collision-avoidance advice, implementers may design descriptions that over-trigger tools in ambiguous contexts, especially for write operations like coupon claiming.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The recommendation to embed user utterances as trigger phrases promotes activation design based on matching informal language, but it does not instruct authors to include counterexamples or ambiguity handling. That is risky in this skill context because the document is a design guide: insecure phrasing patterns may be propagated across many MCP tools, amplifying accidental invocation and unsafe autonomous tool use.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The activation text uses very broad triggers such as “review”, “检查”, and “审查” without narrowing them to MCP tool/interface-design contexts. This can cause the skill to activate for unrelated user requests, leading to incorrect delegation, context hijacking, and reduced reliability of security-sensitive review workflows.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The skill content is written to operate in Chinese and does not offer language negotiation or fallback behavior. In multilingual environments this can cause misunderstanding of findings, incorrect user expectations, or unsafe review outcomes if users assume the skill will respond accurately in another language.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 80% confidence
Finding: The file presents the guidance entirely in Chinese and frames the material around China-specific examples, but it does not state that the language/locale limitation is intentional or optional for users. Under the stated policy, forcing a specific language without opt-in can be a natural-language policy issue unless the locale constraint is clearly documented and justified.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.