The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill includes a placeholder for user-supplied tasks which could potentially be used to deliver indirect instructions.
Ingestion points: Untrusted data enters the agent context via the $ARGUMENTS placeholder in SKILL.md.
Boundary markers: The skill does not use specific delimiters or instructions to treat the $ARGUMENTS content as untrusted.
Capability inventory: Implementation involves the use of Bash, Write, and Edit tools as defined in the skill's frontmatter.
Sanitization: No explicit sanitization or validation of the input provided via $ARGUMENTS is present in the skill instructions.
[EXTERNAL_DOWNLOADS]: The skill's code templates reference the @evenrealities/even_hub_sdk Node.js package, which is the official development kit for the vendor's hardware and platform.

handle-input