Security audit

天气

Security checks across malware telemetry and agentic risk

Overview

This weather skill is not clearly malicious, but it asks for broad read/write/command authority and its documentation is inconsistent and much wider than weather lookup.

Install only after the publisher narrows this skill to weather-specific behavior, removes unrelated automation claims, explains whether an API key is actually required and for which provider, and either removes read/write/exec access or documents tightly scoped, user-directed uses for those tools.

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)

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The documentation materially contradicts itself about whether the skill needs an API key: it advertises 'no API key required' while later instructing users to export API_KEY and stating the feature boundary requires a key. This ambiguity can mislead operators into unsafe setup, unnecessary secret provisioning, or incorrect trust assumptions about how the skill actually fetches data.

Intent-Code Divergence

High

Confidence: 98% confidence
Finding: The skill is labeled as a weather tool, but the documentation claims unrelated design, workflow, file-processing, and other broad automation capabilities. This kind of purpose drift is dangerous because it can justify overbroad permissions and cause an agent or user to invoke the skill in contexts far outside a simple weather lookup, increasing the chance of misuse.

Context-Inappropriate Capability

High

Confidence: 97% confidence
Finding: The declared read and write capabilities are not justified by the stated weather-retrieval purpose. Unnecessary file access expands the blast radius from a simple network/data lookup into potential local data exposure or file modification if the skill is invoked or interpreted too broadly.

Context-Inappropriate Capability

Medium

Confidence: 97% confidence
Finding: The declared read and write capabilities are not justified by the stated weather-retrieval purpose. Unnecessary file access expands the blast radius from a simple network/data lookup into potential local data exposure or file modification if the skill is invoked or interpreted too broadly.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The invocation guidance is overly broad, contradictory, and unrelated to weather, including references to design, branding, and workflow optimization. Ambiguous trigger scope is dangerous in agent systems because it increases accidental activation and can make an over-privileged skill run in contexts where users did not intend a weather tool to operate.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The scenario labels are generic and do not meaningfully constrain what the weather skill is for. In combination with broad tool permissions, vague trigger definitions can cause the agent to route unrelated requests to this skill, increasing the chance of unsafe file or command operations under a misleading weather label.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.