The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses prompt templates that interpolate user-provided strings—such as product descriptions, user profiles, and competitor data—into the agent's instructions. This creates a surface for indirect prompt injection where a malicious user could potentially override the intended logic of the pricing advice prompts.
Ingestion points: Multiple template sections in SKILL.md containing placeholders like [what it does], [who uses it], and [competitor prices].
Boundary markers: The templates do not utilize specific delimiters or instructions to ignore potential commands embedded within the user-provided data.
Capability inventory: The skill does not implement any tools for network access, file system modification, or command execution, limiting the potential impact of an injection.
Sanitization: No input validation or escaping is applied to the user-supplied values before they are placed into the prompts.
[NO_CODE]: The skill consists entirely of instructional Markdown content and prompt templates. It does not include scripts, binaries, or automated task configurations.

pricing