latex-version-control

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No evidence of instructions designed to bypass safety filters or override agent behavior was found. The instructions focus on formatting and data entry requirements.
  • [DATA_EXFILTRATION]: The skill does not perform network requests or access sensitive local files. It operates entirely within the context of generating LaTeX code.
  • [REMOTE_CODE_EXECUTION]: There are no patterns involving the download or execution of remote scripts or binary payloads.
  • [COMMAND_EXECUTION]: The skill does not invoke any system commands or subprocesses. It provides static LaTeX templates.
  • [OBFUSCATION]: The content is provided in plain text and standard LaTeX syntax. No hidden or encoded characters were detected.
  • [INDIRECT_PROMPT_INJECTION]: While the skill ingests user input for version entries (author and changes), it lacks the dangerous capabilities required to facilitate an exploit through this surface.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 07:02 AM