The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run to call the twitter CLI tool. The use of a list for arguments instead of a single string ensures that commands are executed without shell interpretation, effectively preventing command injection attacks.
[EXTERNAL_DOWNLOADS]: The script includes functionality to fetch content from r.jina.ai (Jina Reader) as an optional fallback mechanism. Jina is a recognized service for converting web pages into LLM-friendly formats. The skill requires the agent to obtain explicit user approval before initiating these external network requests.
[CREDENTIALS_UNSAFE]: The skill instructions and code reference the use of environment variables and session cookies for authentication (e.g., TWITTER_AUTH_TOKEN, JINA_API_KEY). The skill includes mandatory instructions for the agent to never print or persist these tokens and to notify the user before any credentialed operation occurs.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests data from external, untrusted sources (X/Twitter posts).
Ingestion points: The fetch_x_content.py script retrieves post text, metadata, and articles from the X platform.
Boundary markers: The skill provides instructions to the agent to distinguish between the fetched content and subsequent analysis, helping to prevent the agent from accidentally following instructions embedded in the tweets.
Capability inventory: The skill possesses the ability to execute the twitter CLI and perform network operations to retrieve web content.
Sanitization: While the skill validates input URLs, it does not sanitize the text content of the posts themselves, relying on the LLM's own safety guardrails and the provided boundary instructions.

x-twitter-reader