用天气可视化你的心情！基于 AI 情绪分析，将文字转化为天气图标 + 治愈文案。

• Outcome: Review
• Latest audit: May 28, 2026 · 11:42 AM UTC
• Version: 1.0.0

Overview

This mood-analysis skill is not clearly malicious, but it needs review because it can route sensitive emotional text through a third-party AI CLI, trigger on broad everyday phrases, and install unpinned external npm code.

Install only if you trust the external mood-weather-cli npm package and are comfortable sending mood text to DeepSeek. Use a dedicated API key, avoid entering highly sensitive personal or medical details, and check whether the CLI stores history or identifiers and how to delete them.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (11)

Missing User Warnings

Medium

• Confidence: 81% confidence
• Finding: The README states that analysis history and mood statistics/reporting are or will be stored, and separately documents a user identifier, but it provides no privacy notice, retention details, consent model, or security expectations. For a mood-analysis tool, stored emotional content can be sensitive personal data, so silent retention increases privacy and misuse risk.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The README instructs users to configure a DeepSeek API key and says the tool is AI-driven, but it does not clearly warn that users' mood text will be sent to an external third-party service for processing. Because mood entries can contain highly sensitive personal information, lack of transparent disclosure can lead to unintended exfiltration of private data off-device.

Vague Triggers

Medium

• Confidence: 89% confidence
• Finding: The trigger phrases are broad, common conversational terms such as '心情怎么样' and '今天天气', making accidental activation likely during normal chat. This is risky because the skill may process sensitive emotional text or invoke external tooling/API calls without clear user intent to use the skill.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: Keyword-based activation without strict scope boundaries means ordinary messages containing emotion-related words can trigger the skill unexpectedly. In this context, accidental activation is more dangerous because the skill may transmit user text to DeepSeek and potentially log or associate it with user identifiers.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill describes analyzing user text but does not warn that the text may be transmitted to an external API for processing. Because emotional disclosures often contain sensitive personal information, omitting this disclosure undermines informed consent and increases privacy risk.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The planned history and logging capabilities describe storing emotional analysis data but provide no user-facing warning about retention, export, or privacy controls. Mood inputs can reveal mental state, relationships, and health-adjacent information, so silent persistence creates a meaningful privacy and misuse risk.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: The trigger list includes broad natural-language phrases such as '分析情绪', '心情怎么样', 'my mood', and 'how am I feeling', which can easily appear in ordinary conversation and unintentionally invoke the exec action. Because the action forwards extracted user text to an external CLI, accidental activation can cause unintended command execution paths, privacy leakage of user text, or confusing behavior in unrelated conversations.

Vague Triggers

High

• Confidence: 98% confidence
• Finding: Including the standalone help trigger '帮助' is extremely collision-prone because it is a common, generic request users may make in many contexts. This can hijack unrelated help requests and route them into this skill's help command, creating denial of intended functionality and unexpected execution of local code via the Node help script.

Ssd 3

Medium

• Confidence: 91% confidence
• Finding: The spec proposes history, statistics, and report export features for mood data without privacy limits, minimization, or purpose restrictions. Retaining and exporting emotional analysis data can expose highly sensitive behavioral patterns, especially if linked across time or shared outside the original context.

Ssd 3

Medium

• Confidence: 95% confidence
• Finding: The skill explicitly describes recording every analysis result, saving history, and associating data with platform-specific user identifiers. This combination materially increases privacy risk because it enables long-term profiling of sensitive emotional state tied to identifiable individuals across channels.

Known Vulnerable Dependency: axios==1.4.0 — 10 advisory(ies): CVE-2025-62718 (Axios has a NO_PROXY Hostname Normalization Bypass that Leads to SSRF); CVE-2026-42044 (Axios: Invisible JSON Response Tampering via Prototype Pollution Gadget in `pars); CVE-2026-25639 (Axios is Vulnerable to Denial of Service via __proto__ Key in mergeConfig) +7 more

High

• Category: Supply Chain
• Confidence: 96% confidence
• Finding: The manifest explicitly includes axios 1.4.0, which is identified as having multiple published security advisories, including SSRF-related and prototype-pollution/DoS issues. In a CLI skill that likely performs remote API calls for mood analysis, a vulnerable HTTP client can materially increase risk if attacker-controlled inputs, proxy settings, or response parsing paths are reachable.

Static analysis

No suspicious patterns detected.