writer

Pass

Audited by Gen Agent Trust Hub on Jun 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses subprocess.run to execute system-level commands, including unzip, zip, and soffice (LibreOffice), for file manipulation and document conversion.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection vulnerability surface within the generate_from_template Python logic.
  • Ingestion points: Untrusted data enters the skill via the variables dictionary used in the generate_from_template function in SKILL.md.
  • Boundary markers: There are no delimiters or instructions used to separate the template's XML structure from the interpolated variable content.
  • Capability inventory: The skill can execute shell commands (unzip, zip, soffice) and perform file system writes.
  • Sanitization: The skill uses basic string replacement (content.replace) without escaping or validating inputs, which could allow malicious instructions inside the variables to manipulate the document structure or influence downstream agents.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 26, 2026, 02:40 PM
Security Audit — agent-trust-hub — writer