Back to skill

Security audit

Research skill

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward academic-paper writing helper, with file, web-search, and LaTeX compile behavior that fits its stated purpose.

Install only if you are comfortable letting the agent read the selected project directory, generate paper files under the chosen output path, use web search for references, and optionally run the disclosed LaTeX compile command. For confidential or unpublished research, ask the agent to avoid web search or summarize/redact sensitive details first.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The entire skill description, instructions, examples, and output template are written in Chinese, and there is no indication that users may choose another language. This can violate language/locale policy when a skill implicitly enforces a specific language without opt-in or justification.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill is designed to read project code and experiment results and to use web search for references, but it does not warn users about the privacy and data-handling implications of accessing potentially sensitive local materials and sending derived context to external services. In research settings, source code, unpublished results, and manuscript content may be confidential, making undisclosed data access and external lookup risky.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill explicitly writes multiple LaTeX and bibliography files to disk, but the metadata and user-facing description do not clearly warn about these filesystem side effects. This can surprise users into modifying repositories or local directories they did not intend to change, especially when output paths may overlap with existing project files.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill advertises use of Bash to compile the generated paper and even provides a pdflatex/bibtex command sequence, but it does not prominently warn users that shell commands may be executed. Shell execution materially raises risk because it can consume resources, invoke local tooling on attacker-influenced inputs, and potentially trigger unsafe behavior in LaTeX toolchains or build scripts.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.