skills/openclaw/skills/svg-draw/Gen Agent Trust Hub

svg-draw

Pass

Audited by Gen Agent Trust Hub on Feb 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • Command Execution (LOW): The skill utilizes a bash script scripts/svg_to_png.sh to execute the system utility rsvg-convert. While the script uses double quotes for variables, it does not perform type or value validation on the WIDTH, HEIGHT, or path arguments. This could allow an attacker to inject additional command-line flags if the agent is tricked into passing malicious strings.
  • Indirect Prompt Injection (LOW): The skill processes SVG files and command arguments which constitute an attack surface for indirect injection if the source data is untrusted.
  • Ingestion points: SVG file content and shell script arguments in scripts/svg_to_png.sh.
  • Boundary markers: Absent; the skill does not use delimiters or instructions to ignore embedded XML directives in the SVG source.
  • Capability inventory: Execution of rsvg-convert and filesystem write access as defined in scripts/svg_to_png.sh.
  • Sanitization: Absent; there is no validation of the SVG schema or sanitization of the dimensions before they are passed to the system renderer.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 19, 2026, 11:44 PM