The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data which could potentially contain malicious instructions intended to influence agent behavior.
Ingestion points: The skill handles user-provided inputs including prompt, image_url, video_url, and callback_url parameters within the SKILL.md instructions.
Boundary markers: There are no explicit delimiters or structural markers defined to separate user-provided content from agent instructions in the prompt construction guidance.
Capability inventory: The skill enables the agent to perform network requests to the Kling AI API (api-singapore.klingai.com) and access local files for conversion to Base64 (e.g., image_path).
Sanitization: The skill provides proactive security instructions to validate that callback_url parameters are absolute http or https URLs before use.

kling-official