The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the shell interpreter to execute internal Python scripts for lifecycle management, including initialization (init_skill.py), validation (quick_validate.py), and packaging (package_skill.py). These scripts perform localized filesystem operations such as directory creation and zipping files.
[COMMAND_EXECUTION]: The workflow includes instructions for the agent to generate and test Python scripts in the local environment, executing them via the shell interpreter to verify functionality.
[PROMPT_INJECTION]: The skill's role as a template generator creates an indirect prompt injection surface through the ingestion of user-provided requirements.
Ingestion points: User-provided descriptions and examples are ingested in Step 1 and Step 2 of the workflow outlined in SKILL.md.
Boundary markers: There are no explicit delimiters or instruction-guarding markers used to separate user-supplied content when it is interpolated into generated skill components.
Capability inventory: The tool has access to file system writes and command execution via shell tools.
Sanitization: Metadata format and length are validated by the package scripts, but the logic of generated Python scripts is not subjected to automated security sanitization.

skill-creator