The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input via prompts and external media URLs (images, videos, and audio) for processing by the Kling API.
Ingestion points: The prompt, start_image_url, end_image_url, video_url, audio_url, and text parameters in API requests.
Boundary markers: None specified in the instructions.
Capability inventory: Performs network operations via curl to api.acedata.cloud.
Sanitization: Not explicitly implemented in the provided examples.
[EXTERNAL_DOWNLOADS]: The skill documentation includes examples that connect to https://api.acedata.cloud, which is the legitimate API endpoint for the skill vendor, AceDataCloud.
[CREDENTIALS_UNSAFE]: The skill requires an ACEDATACLOUD_API_TOKEN. It correctly advises the user to manage this secret using a .env file and explicitly warns against committing the key or including it in prompts.

kling-video