The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill accepts external input through the $0 and $ARGUMENTS parameters to generate UX microcopy. Because it lacks boundary markers or sanitization instructions, a malicious user could provide input designed to override the skill's instructions or manipulate the agent into using its tools for unintended purposes.
Ingestion points: User-provided context and description fields mapped to $0 and $ARGUMENTS in SKILL.md.
Boundary markers: Absent. The skill does not instruct the agent to use specific delimiters or to ignore instructions embedded within the provided context or description.
Capability inventory: The skill is configured with Read, Write, Edit, Grep, and Glob tools, enabling file system modification.
Sanitization: Absent. No filtering, validation, or escaping is performed on the input before it is processed by the agent.

ss-copy