The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes dynamic context injection to provide environment information during initialization.
Evidence: Usage of !date and !node --version in SKILL.md to display timestamps and versioning.
Note: These commands are benign and used for informational purposes only.
[PROMPT_INJECTION]: The skill includes an attack surface for indirect prompt injection through the use of user-supplied arguments.
Ingestion points: The $ARGUMENTS placeholder is used multiple times in SKILL.md to define optimization targets.
Boundary markers: Absent; user input is interpolated directly into the instruction flow without delimiters or safety warnings.
Capability inventory: The skill directs the agent to modify files, optimize site structure, and generate documentation, which involves file-system writes and reads.
Sanitization: Absent; there is no instruction for the agent to validate or escape content within $ARGUMENTS.

seo