The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to install the hermes-tweet plugin directly from the developer's GitHub repository (Xquik-dev/hermes-tweet). This is a standard setup procedure for the skill's functionality.
[PROMPT_INJECTION]: The skill is designed to process untrusted data from an external source (X/Twitter tweets and replies), which presents a surface for indirect prompt injection. This risk is classified as safe due to the implementation of strict human-in-the-loop controls for all write-capable actions.
Ingestion points: Untrusted content is ingested via the tweet_read tool from X/Twitter search results and reply threads (identified in SKILL.md).
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions for the data retrieved from X/Twitter.
Capability inventory: The skill utilizes the tweet_action tool for posting tweets, replies, and direct messages (identified in SKILL.md).
Sanitization: Risk is mitigated by a mandatory safety protocol requiring the agent to present exact drafts to the user and obtain explicit approval before executing any write operations.
[SAFE]: The skill adheres to several security best practices, such as separating read and write capabilities, recommending the use of environment variables for API keys (XQUIK_API_KEY), and requiring auditable logs and source links for all automated reports.

x-twitter-automation