Skills
skills/haboshi/claude-code-skills/line-sticker-creator/Gen Agent Trust Hub

line-sticker-creator

Warn

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the find command to dynamically locate dependency scripts in the user's plugin cache directory.
  • [COMMAND_EXECUTION]: It executes local Python scripts using uv run with dynamically computed file paths, which could be exploited if the plugin directory is compromised.
  • [PROMPT_INJECTION]: User-provided character descriptions and themes are interpolated into image generation prompts without sanitization or boundary markers, creating a surface for indirect prompt injection.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 12, 2026, 03:42 AM