The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The SKILL.md guide recommends a multi-agent acceleration pattern that interpolates directory paths into agent instructions.
Ingestion points: File and directory names from the filesystem are ingested via DirectoryScanner in diskcleaner/core/scanner.py.
Boundary markers: No explicit delimiters or warnings to ignore embedded instructions are shown in the SKILL.md code examples.
Capability inventory: The skill has capabilities to create new agents and execute subprocesses (e.g., subprocess.run in diskcleaner/core/process_manager.py).
Sanitization: No sanitization or escaping of the interpolated path data is provided in the documentation's code examples.
[DYNAMIC_EXECUTION]: Internal diagnostic and bootstrap scripts use dynamic execution techniques for environment verification and module discovery.
scripts/check_skill.py uses compile() and __import__() to verify the syntax and availability of the skill's own local modules.
scripts/skill_bootstrap.py uses __import__() to dynamically locate and load the diskcleaner core package across various possible installation paths.

disk-cleaner