Security audit

Academic Research Skip

Security checks across malware telemetry and agentic risk

Overview

This skill is a straightforward academic search and literature-review helper, with expected external API calls and local output/cache behavior that users should understand.

Install only if you are comfortable with academic queries, author names, and DOIs being sent to OpenAlex or Unpaywall. Avoid confidential research topics, choose --output paths carefully because files may be overwritten, and clear /tmp/litreview_cache if cached paper metadata should not remain on the machine.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (6)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 92% confidence
Finding: The skill advertises and instructs use of scripts that perform network access, read data, and write output/cache files, but the skill metadata does not declare corresponding permissions. This mismatch weakens review and runtime trust boundaries because a caller or platform may not realize the skill can access external resources and write local files such as /tmp caches or output reports.

Context-Inappropriate Capability

Medium

Confidence: 91% confidence
Finding: The script persists remote API responses under /tmp and can also write generated review output to an arbitrary user-specified path. While this supports normal functionality, it expands the skill from transient research/search into local data persistence, which can expose queried topics, retrieved metadata, and generated content to other local users or overwrite files if invoked unsafely.

Vague Triggers

Medium

Confidence: 82% confidence
Finding: The trigger description is broad enough to match generic requests about research, synthesis, or reference gathering, which can cause the skill to activate in contexts broader than intended. Because the skill has network and file-write behavior, over-broad invocation increases the chance of unnecessary external calls, unintended data handling, or use in contexts where academic web search is not appropriate.

Missing User Warnings

Low

Confidence: 88% confidence
Finding: User-supplied research topics are transmitted to the OpenAlex API, which creates a privacy and data-sharing risk if users enter confidential, proprietary, or sensitive topics. In the context of an academic research skill this network behavior is expected, but the lack of explicit disclosure means users may not realize their query leaves the local environment.

Missing User Warnings

Low

Confidence: 93% confidence
Finding: The script writes to any path supplied via --output using Path(args.output).write_text(output) without warning or overwrite protection. This can unintentionally clobber existing files or place generated content in unexpected locations, especially when the tool is called by another agent on behalf of a user.

Missing User Warnings

Low

Confidence: 85% confidence
Finding: The script automatically attaches a hardcoded personal contact email to all API requests, disclosing operator-identifying information to third parties without runtime consent or configuration. This can leak maintainer identity, create unwanted correlation across users and searches, and expose the address to logging, rate-limit, or abuse workflows at external services.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.