Scisurvey

• Outcome: Pass
• Latest audit: Jun 4, 2026, 8:44 AM
• Version:
  1.0.0

Overview

SciSurvey is a coherent academic review skill; its file creation and optional document conversion are purpose-aligned but should be understood before use.

Install this if you want an automated Sciverse-based literature review workflow. Before requesting DOCX, PDF, or LaTeX output, be aware it may create files in the working directory and may run local conversion tools if available; use Markdown output if you want the lowest-side-effect path.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Context-Inappropriate Capability

Medium

• Confidence: 95% confidence
• Finding: The skill instructs the agent to run local shell commands and invoke external tooling such as pandoc, xelatex, pdflatex, bibtex, brew, and apt. That expands the trust boundary from Sciverse-only retrieval into host-side command execution and file generation, which can modify the local environment or trigger unintended subprocess activity without clear user consent or capability checks.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The description does not clearly define when the skill should not run, so ordinary writing or research-assistance requests may be misrouted into this autonomous survey pipeline. In this context, ambiguity is riskier because the skill is designed to perform many actions and maintain a long procedural flow, making unintentional invocation more operationally dangerous than a simple read-only helper.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The description does not clearly define when the skill should not run, so ordinary writing or research-assistance requests may be misrouted into this autonomous survey pipeline. In this context, ambiguity is riskier because the skill is designed to perform many actions and maintain a long procedural flow, making unintentional invocation more operationally dangerous than a simple read-only helper.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The skill includes instructions to create files and potentially invoke local compilation/conversion tools, but the user-facing description does not warn about these side effects. This undermines informed consent and can surprise users with filesystem writes or local tool execution that exceeds expectations for a literature-review assistant.

VirusTotal

57/57 vendors flagged this skill as clean.