The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/init_skill.py script automates the creation of skill directory structures and template files. It applies executable permissions (chmod 0o755) to generated example scripts, which is standard behavior for a project initialization tool.
[COMMAND_EXECUTION]: The scripts/package_skill.py script reads the contents of a skill folder to create a compressed .skill archive. It includes robust security checks, specifically a path validation function (_is_within) to ensure all files being packaged are within the skill's root directory, and an explicit check to skip symlinks.
[PROMPT_INJECTION]: The skill's primary function includes auditing and improving other skills, which involves ingesting and processing external instruction files. This creates a surface for indirect prompt injection.
Ingestion points: Reads and processes the content of SKILL.md files and other resources within a skill directory being audited or improved.
Boundary markers: The instructions do not define specific delimiters for separating the user-provided skill content from the auditing agent's logic.
Capability inventory: The skill has the capability to write files, modify permissions (scripts/init_skill.py), and package files into archives (scripts/package_skill.py).
Sanitization: The init_skill.py script uses regex normalization for skill names to prevent injection via filenames, and package_skill.py uses path resolution to prevent traversal attacks.

skill-creator