Back to skill

Security audit

Thesis Review Pro

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed thesis-review helper that reads user-provided papers and generates Chinese review documents, with privacy and storage cautions but no hidden execution or malicious behavior found.

Install only if you are comfortable with the agent reading thesis PDFs, extracting author and academic-profile details, searching the web for comparison literature, and writing review .docx files. For blind review or confidential theses, remove identifying pages first or request anonymized filenames and outputs.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The description states the skill is for Chinese-language thesis review, and the template later hardcodes Chinese review content and Chinese font requirements. There is no indication that users may choose another language or that the locale restriction is required for a specific regional compliance context.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The manifest lists generic triggers such as "论文评审", "评审意见", and "毕业论文" without constraints, negative examples, or activation boundaries. These phrases could match ordinary discussion about theses or review comments rather than an explicit request to invoke this skill.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill explicitly directs extraction of personal and academic information from uploaded PDFs, including author, school, project history, and achievements, but provides no privacy notice, consent boundary, minimization guidance, or retention limits. In a thesis-review context, this can lead to unnecessary processing of sensitive personal data and accidental disclosure in downstream outputs or logs.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs creation of .docx files containing review content and personal identifiers such as the author's name, with a filename pattern that embeds identity data, but gives no warning about local file creation, storage location, or access control. This increases the risk of inadvertent data exposure through shared folders, backups, sync services, or predictable archived filenames.

Content

No source excerpt is available for this finding.

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
81% confidence
Finding

The skill enumerates internal-looking paths and historical artifact names, including a memory file containing '13篇论文评审记录' and archived review documents. In context, this can leak implementation structure and signal the existence of potentially sensitive retained review data, which may aid unauthorized discovery or targeted retrieval attempts in systems that expose tool/file namespaces.

Content

Scanner excerpt · SKILL.md (reported line 240)May include surrounding context.

md
| 文件 | 说明 |
|------|------|
| `skills/thesis-review-pro/SKILL.md` | 本技能文档 |
| `memory/2026-05-21.md` | 13篇论文评审记录 |
| `评审意见_*.docx` | 历史评审意见存档 |

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The package description is entirely in Chinese and describes the skill as a Chinese academic thesis review tool, but it does not state that the language/locale is optional or limited to a justified region-specific context. Under the language/locale policy, natural-language content that effectively constrains usage to a specific language without opt-in can be a policy violation.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.