logo-batch-generator
Warn
Audited by Gen Agent Trust Hub on Jun 24, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Risk of command injection through unvalidated user input. The skill's instructions guide the agent to construct shell commands in SKILL.md using brand names provided by the user as part of the --output path argument. Without explicit sanitization, shell metacharacters in the brand name (e.g., semicolons, backticks) could lead to arbitrary command execution on the host system.\n- [COMMAND_EXECUTION]: Disclosure of host environment metadata. The hardcoded examples in SKILL.md use absolute file paths (e.g., /Users/ugreen/Documents/obsidian/), revealing the host's username and internal directory structure.\n- [PROMPT_INJECTION]: Vulnerability to indirect prompt injection through unvalidated data ingestion.\n
- Ingestion points: The skill processes user-supplied brand descriptions and content extracted from user-uploaded images in SKILL.md (Phase 1 and 2).\n
- Boundary markers: No delimiters or safety instructions are used when interpolating user data into prompt templates in references/logo-styles.md.\n
- Capability inventory: The skill executes a local Python script (scripts/generate_image.py) with the capability to create directories and write files to the filesystem.\n
- Sanitization: No sanitization or validation logic is defined for user-controlled strings before they are used as command-line arguments.
Audit Metadata