The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by passing user-provided arguments directly to sub-skills.
Ingestion points: User input is ingested via the {{args}} parameter in SKILL.md.
Boundary markers: There are no explicit delimiters (e.g., XML tags or triple quotes) used to isolate the user input from the routing instructions.
Capability inventory: The skill's environment allows access to tools such as Bash, Write, Edit, WebFetch, and WebSearch as specified in the allowed-tools metadata.
Sanitization: The skill does not perform pre-processing or validation on the {{args}} content before passing it to the Skill tool.

helm