The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by interpolating untrusted user input ({{args}}) directly into the agent's instruction context.
Ingestion points: The {{args}} placeholder in SKILL.md is replaced with user-supplied data at runtime.
Boundary markers: No delimiters (e.g., triple quotes, XML tags) or instructions to ignore embedded commands are present to isolate the user input from the skill's instructions.
Capability inventory: The skill utilizes tools such as Bash, Write, Edit, and WebFetch, which represent a significant capability tier if an injection is successful.
Sanitization: There is no evidence of input validation or sanitization for the interpolated data.

warden