skill-creator
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEPRIVILEGE_ESCALATIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The
scripts/init_skill.pyscript applies execution permissions (chmod 0o755) to a dynamically generated example Python script (scripts/example.py). - [DYNAMIC_EXECUTION]: The
scripts/init_skill.pyscript generates new Python source files by interpolating user-provided inputs into embedded string templates during the skill initialization process. - [INDIRECT_PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection as it accepts unvalidated user input via command-line arguments to create and name file system directories and files.
- Ingestion points: Command-line arguments
skill-nameand--pathprocessed inscripts/init_skill.pyandscripts/package_skill.py. - Boundary markers: None present in the generated SKILL.md template or bundled resources.
- Capability inventory: File system operations including directory creation (
mkdir), file writing (write_text), permission modification (chmod), and zip archive creation (ZipFile). - Sanitization: Basic validation in
scripts/init_skill.pyandscripts/quick_validate.pyensures skill names conform to a lowercase hyphenated format. - [COMMAND_EXECUTION]: The skill instructions direct the agent to execute local Python scripts (
init_skill.py,package_skill.py,quick_validate.py) to manage the skill lifecycle, which involves reading and writing to the local file system.
Audit Metadata