julia-vibe-coding

Fail

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The setup instructions in references/julia_mcp_setup.md require cloning and executing code from an unverified third-party repository (github.com/aplavin/julia-mcp). It specifically instructs the user to run a Python server (server.py) from this untrusted source using uv run to enable the skill's functionality.
  • [EXTERNAL_DOWNLOADS]: The skill relies on software hosted on a personal GitHub account (aplavin) that does not belong to a trusted organization or well-known service, bypassing standard security verification for dependencies.
  • [COMMAND_EXECUTION]: The skill provides the agent with the julia_eval tool, designed to execute arbitrary Julia code. This capability provides a direct path for the agent to perform unauthorized system operations or access sensitive data if it is misdirected by malicious input.
  • [PROMPT_INJECTION]: The workflow involves reading and executing local files (e.g., using include on files in the scripts/ directory). This creates a surface for indirect prompt injection where an attacker could place malicious Julia code in a project file that the agent then unknowingly executes. The skill lacks sanitization or boundary markers to prevent the execution of embedded instructions in these data sources.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 7, 2026, 11:09 AM