Back to skill

Security audit

oldglycine-paper-reference-adder

Security checks for vulnerabilities and agentic risk

Overview

The skill has a coherent citation-assistance purpose, but its included script can rewrite paper files without the promised backup and can overwrite files outside the project through an unchecked report path.

Review before installing or running. Use version-pinned dependencies in an isolated environment, avoid the documented pdflatex pip package unless independently verified, keep the report path relative and inside the paper project, and make your own backup or use version control before allowing the skill to edit TeX/BibTeX files.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T09 · Insecure Skill Coding Practices

Warning
Location
main.py:95
Finding

Unchecked Report Path Allows Arbitrary File Overwrite

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
reference/CODE.MD:6
Finding

Unpinned and Unnecessary Third-Party Package Installation

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Note
Location
main.py:77
Finding

TeX Files Are Rewritten Without the Documented Backup Safeguard

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (10)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The natural-language instructions and usage details are presented only in Chinese, and there is no indication that users can opt into another language or that the skill is intended solely for a Chinese-language audience. SQP-3 applies to language or locale policy violations when a specific language is effectively forced without user opt-in.

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 performs write operations against TeX/BibTeX project files by adding BibTeX entries and inserting citation markers, but the description does not clearly warn users that source files will be modified. This can lead to unintended document changes, repository pollution, or overwriting user work, especially in collaborative Overleaf-style projects where file edits have immediate effect.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The workflow sends content-derived keywords from the user's paper to Google Scholar via the scholarly library, but the skill description does not disclose this network exfiltration/privacy behavior. Even if only keywords are sent, they may reveal unpublished research topics, sensitive project themes, or confidential manuscript content to third-party services.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The top-level description says the tool will add BibTeX entries and insert citation markers into relevant sections of the TeX project. In the code, there is no logic that edits any .tex file or inserts \cite commands; it only reads TeX content, searches Scholar, appends entries to custom.bib, and writes ADD.MD.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The tool derives keywords from the user's paper and sends them to Google Scholar via the scholarly library, but the description does not warn about external network access or disclosure of paper-derived queries. In research workflows, even partial topic leakage can expose confidential project direction, unpublished work, or sensitive subject matter to third-party services.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The tool derives keywords from the user's paper and sends them to Google Scholar via the scholarly library, but the description does not warn about external network access or disclosure of paper-derived queries. In research workflows, even partial topic leakage can expose confidential project direction, unpublished work, or sensitive subject matter to third-party services.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The docstring states that if no anchor substring is provided, the sentence is appended to the end of the file. However, when the target .tex file is empty, insert_idx becomes 0 and the loop over existing lines never runs, so no sentence is inserted before the file is rewritten. This is a direct contradiction between the documented behavior and actual behavior for that case.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The entire skill documentation is written in Chinese, and there is no indication that language selection is optional or that the skill is intentionally limited to a Chinese-speaking context. This can violate language/locale policy when a specific language is imposed without opt-in.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

The example shows running the script with --project_dir, --num_references, and --output_report arguments. In the code, main() hardcodes these values and does not use argparse or otherwise consume command-line options, so the documented invocation does not match the implementation.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

The documentation states that a detailed report of all added citations will be generated. However, the report generation code only writes a single line containing the total number of citations added, without listing which citations were added or any details about them.

Content

No source excerpt is available for this finding.