The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands (head, grep, ls, cat) to inspect file contents and directory structures. While these are used for information gathering to support code explanation, they represent a surface for shell interaction based on user-provided arguments ($FILE, $NAME, $DIR).
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data (external file contents) through read operations.
Ingestion points: File content is read into the agent's context in SKILL.md using head, grep, and cat commands.
Boundary markers: The instructions do not define clear delimiters or provide explicit warnings to the agent to ignore instructions embedded within the code (e.g., malicious instructions hidden in code comments).
Capability inventory: The skill has read-only capabilities including head, grep, ls, and cat.
Sanitization: There is no evidence of sanitization or filtering of the file content before it is processed by the LLM.

explain