character-sprite
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use standard system commands (
cp,cd) and the ImageMagick suite (magick) to manage, validate, and process generated image assets. - [COMMAND_EXECUTION]: The workflow involves executing a local shell script (
remove_magenta.sh) residing in a specific project directory (/Users/probello/Repos/claude-office/.claude/skills/shared/scripts/) to automate complex background removal tasks. - [INDIRECT_PROMPT_INJECTION]: The skill has an injection surface where user-provided descriptions for characters and clothing are directly interpolated into image generation prompts and used to construct local file paths.
- Ingestion points:
SKILL.mdStep 1 and Step 3. - Boundary markers: None explicitly defined within the prompt templates to isolate user input from the rest of the generation instructions.
- Capability inventory: Image generation tool access, local file system manipulation, and shell command execution.
- Sanitization: No sanitization or validation of the character/clothing description strings is performed before they are processed by the agent.
Audit Metadata