The Agent Skills Directory

[SAFE]: The skill implements a mandatory human-in-the-loop validation step ('Approval Gate') in the workflow. It explicitly states that no content will be published to the Threads Graph API without the user's prior approval of the draft.
[SAFE]: Management of sensitive data, such as the THREADS_ACCESS_TOKEN, is handled through external environment variables as specified in the configuration instructions. This adheres to industry best practices for secret management by avoiding hardcoded credentials.
[PROMPT_INJECTION]: The skill processes untrusted input from external sources (such as blog posts or news summaries) to generate content drafts. While this exposure constitutes a surface for indirect prompt injection, the risk is neutralized by the manual approval requirement and the absence of autonomous execution capabilities.

threads-post-draft