The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes utility scripts (scripts/init_skill.py and scripts/package_skill.py) that perform file system operations such as directory creation, file writing, and recursive file scanning for ZIP archiving. While these are legitimate functions for a scaffolding tool, they involve direct interaction with the local environment.
[PRIVILEGE_ESCALATION]: The scripts/init_skill.py script programmatically sets execution permissions using chmod(0o755) on a dynamically created template file (scripts/example.py). Modifying file permissions on generated content is a sensitive operation, though it is consistent with the skill's purpose of creating functional script directories.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation of new AI instructions based on user-provided descriptions. This creates a surface where malicious or adversarial input from a user could be incorporated into a generated SKILL.md file, potentially affecting other agents that load the resulting skill.
Ingestion points: User-provided descriptions, functionality requirements, and skill names (handled in SKILL.md and scripts/init_skill.py).
Boundary markers: The generation process lacks explicit delimiters or warnings to isolate user-provided content from the structural instructions in the generated templates.
Capability inventory: The skill has the capability to write and modify files on the local disk through its bundled Python scripts.
Sanitization: The scripts/quick_validate.py script provides basic regex validation for skill names and checks for angle brackets in descriptions, but it does not perform deep sanitization of the instructional content being generated.

skill-creator