The Agent Skills Directory

[SAFE]: The skill implements a mandatory human-in-the-loop approval gate. It explicitly instructs the agent to show the draft (caption and media) to the user and requires explicit approval before calling any publishing tools.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied topics and style profiles retrieved via threads_style_load (Ingestion points). The risk of malicious content injection is mitigated by the mandatory human-in-the-loop approval step (Sanitization) which requires users to review the final draft before publication. The skill interacts with the Instagram Graph API using instagram_publish_image, instagram_publish_video, and instagram_publish_reel (Capability inventory). No specific delimiters are used to wrap the input data in the instructions (Boundary markers).
[SAFE]: Instructions for setting up API credentials (IG_ACCESS_TOKEN, IG_USER_ID) are provided for the user to configure their environment. These are standard setup procedures and do not involve hardcoded secrets or automated exfiltration mechanisms.

instagram-post