Security audit

Trending Feed Skill Free

Security checks across malware telemetry and agentic risk

Overview

This skill mainly guides an agent to fetch public GitHub Trending data, with some broad save/export language users should treat carefully.

Install only if you want an agent to run a local Python-based GitHub Trending fetch workflow. Keep writes user-directed: specify the output path yourself, avoid overwriting existing files, and do not provide a GitHub token unless you specifically need higher API limits.

SkillSpector

By NVIDIA

Vulnerability Patterns

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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Description-Behavior Mismatch

Medium

Confidence: 87% confidence
Finding: The skill is presented as a GitHub Trending fetcher, but the documentation explicitly expands behavior into save/export/convert operations and local persistence. In an agent environment with declared write and exec tools, this broadens the operational scope and can cause the agent to create or modify local files beyond a user's likely expectation, increasing the chance of unintended side effects.

Vague Triggers

Medium

Confidence: 78% confidence
Finding: The activation and usage description is overly broad, including generic analytics, reporting, and visualization scenarios not tightly tied to a simple trending fetch task. In agent systems, vague invocation cues can cause the skill to trigger in unrelated contexts, leading to unnecessary use of exec/write-capable tooling and expanding exposure surface.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The capability-scope keyword list is excessively broad and contains many generic terms that could match loosely related user requests. Because the skill also advertises exec and write-capable operations, overbroad routing cues increase the chance an agent invokes this skill when simpler, safer handling would suffice.

Missing User Warnings

Low

Confidence: 81% confidence
Finding: The markdown describes export/save behavior without a clear warning that local files may be created or overwritten. In an agent context, undocumented persistence can surprise users and result in unintended data writes, especially when the skill metadata includes write access.

Missing User Warnings

Low

Confidence: 88% confidence
Finding: Recommending shell redirection to a local file without warning about file creation or overwrite encourages persistence side effects without explicit consent or safeguards. While the example is not malicious, it normalizes writing to disk in a skill whose primary purpose is read-oriented data retrieval.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.