Security audit

API开发工具（免费版）

Security checks across malware telemetry and agentic risk

Overview

This skill is a broadly scoped API development helper that discloses read and command execution access, with no evidence of hidden, destructive, or deceptive behavior.

Install this only if you want an agent to help with API development and are comfortable with it reading project files and running development commands. Review any command, dependency install, file write, deployment, or external service call before allowing it, especially when secrets or production systems are involved.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (6)

Description-Behavior Mismatch

Medium

Confidence: 91% confidence
Finding: The skill’s documented capabilities expand from API lifecycle assistance into broad file processing, information retrieval, and external service use. This scope drift increases the chance the agent will select and use the skill in contexts unrelated to API development, exposing read/exec functionality in situations users may not expect.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: Advertising generic command execution is dangerous because the skill is allowed to use exec and the documentation does not meaningfully constrain what commands may be run. In practice, this can lead to arbitrary local command execution during broad development requests, increasing the risk of destructive actions, data exposure, or unintended system changes.

Context-Inappropriate Capability

Low

Confidence: 78% confidence
Finding: The skill presents itself as able to integrate with external services broadly, but its manifest and documentation do not clearly bound which services, what data may be sent, or under what approval model. This creates risk of unintended data exfiltration or use of third-party endpoints beyond the user’s expectations.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The invocation description is so broad that the skill may be auto-selected for many ordinary programming or development tasks. Because the skill has read and exec permissions, overbroad routing materially increases the chance of unnecessary command execution or file access in sessions where a narrower skill would be safer.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The example triggers are natural-language prompts that are broad and unconstrained, encouraging invocation for large classes of requests without clarifying limits or safety checks. In an agent ecosystem, this increases accidental selection and can expose read/exec capabilities during loosely specified requests.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill description does not clearly warn users that it may execute system commands and modify files, despite having exec capability and discussing command execution elsewhere. This undermines informed consent and can lead users to invoke the skill without understanding the operational and security consequences.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.