The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes Python utility scripts (init_skill.py, generate_openai_yaml.py, quick_validate.py) intended to be executed by the agent to automate the setup of new skill directories, generation of metadata, and validation of skill structure.
[EXTERNAL_DOWNLOADS]: The scripts depend on the PyYAML library for parsing and writing YAML frontmatter and configuration files.
[DYNAMIC_EXECUTION]: The init_skill.py script generates template files, including a placeholder Python script (example.py), to assist in the initial development of new skills.
[PRIVILEGE_ESCALATION]: The initialization script uses chmod(0o755) to ensure that generated example scripts are executable by the user, which is standard behavior for a development tool.
[INDIRECT_PROMPT_INJECTION]: The skill processes user input to populate templates for new skills, representing a potential surface for injection.
Ingestion points: Skill names, interface overrides, and descriptions provided as command-line arguments to init_skill.py and generate_openai_yaml.py.
Boundary markers: The scripts use fixed templates and structured YAML formatting to contain user-supplied strings.
Capability inventory: The skill performs directory creation, file writing, and permission modifications on the local filesystem.
Sanitization: The normalize_skill_name function uses regular expressions to restrict skill names to alphanumeric characters and hyphens, and yaml_quote is employed to escape strings in generated configuration files.

skill-creator