Security audit

代码分析工具专业版

Security checks across malware telemetry and agentic risk

Overview

The skill is a code-analysis helper, but it asks for broad agent powers while mixing local analysis, automatic history storage, callbacks, and external API use without clear boundaries.

Review this before installing in repositories with proprietary code or secrets. Use it only when you explicitly want repository analysis, keep network/API and callback features disabled unless needed, verify destination URLs and environment variables, and periodically purge .code-analyze history if reports may contain sensitive content.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Intent-Code Divergence

Medium

Confidence: 87% confidence
Finding: The documentation gives conflicting statements about whether the skill works offline, whether it needs network access, and whether an API key is required. In a skill with Bash/Write/Edit capabilities, this ambiguity can cause operators or downstream agents to make unsafe assumptions and unintentionally transmit code or secrets to external services.

Intent-Code Divergence

Medium

Confidence: 85% confidence
Finding: The Q&A instructs users to configure environment variables and API keys even though earlier sections say no extra API key is needed. This inconsistency can lead users to provision credentials unnecessarily or expose secrets in environments where they believed the tool was purely local.

Vague Triggers

High

Confidence: 93% confidence
Finding: The activation description is extremely broad, covering generic coding, debugging, testing, and deployment scenarios rather than a narrowly scoped analysis function. Overbroad triggers increase the chance that an agent invokes a powerful skill in unrelated contexts, expanding access to Read/Write/Edit/Bash beyond user intent.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The supported-keyword list is excessively broad and ambiguous, making activation boundaries unclear. In agent systems, fuzzy routing can cause this skill to intercept benign requests and perform unintended file access, command execution, or archival behaviors.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The skill describes automatic archival and trend tracking of analysis history without a prominent consent and retention warning. Because analyzed content may include proprietary code, secrets, or sensitive findings, silent local persistence increases confidentiality and compliance risk.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill supports callback URLs and external API/webhook usage without a prominent warning that analysis inputs and results may leave the local system. For a code-analysis skill, this is especially risky because source code, vulnerability details, and embedded secrets could be exfiltrated to third-party endpoints.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.