The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/init_skill.py generates an executable boilerplate script (example.py) by injecting the user-supplied skill_name directly into a Python template via string formatting. This lack of sanitization allows for arbitrary code injection into the generated file if a malicious skill name is used.- [COMMAND_EXECUTION]: The scripts/init_skill.py script programmatically modifies file permissions to executable (chmod 755) for the boilerplate script it generates.- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes user-provided skill files during validation and packaging.
Ingestion points: scripts/quick_validate.py and scripts/package_skill.py read and parse SKILL.md and other directory contents.
Boundary markers: The skill does not use protective delimiters or specific instructions to ignore embedded prompts within the files it processes.
Capability inventory: The toolset has the capability to create directories, write files, modify permissions, and create ZIP archives.
Sanitization: Validation is limited to regex-based format checks for specific metadata fields and does not include sanitization of the file bodies.- [COMMAND_EXECUTION]: The scripts scripts/init_skill.py and scripts/package_skill.py accept arbitrary file paths from command-line arguments and perform file system operations (read, write, mkdir) at those locations without strict path validation or restriction.

skill-creator