The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from the X/Twitter platform, creating a surface for indirect prompt injection attacks where malicious instructions hidden in posts or messages could influence the agent's behavior.
Ingestion points: The agent retrieves content from the X platform through actions such as search_posts, lookup_timeline, lookup_dms, and lookup_news as defined in SKILL.md.
Boundary markers: The instructions do not provide specific delimiters or warnings to help the agent distinguish between its system instructions and the content retrieved from X.
Capability inventory: The skill provides extensive write capabilities, including posting tweets (create_post), sending direct messages (send_dm), and making network fetches (upload_media).
Sanitization: No sanitization or validation of the retrieved social media content is specified before the agent processes it.
[EXTERNAL_DOWNLOADS]: The upload_media action includes a source_url parameter that allows the tool to fetch media content from arbitrary public URLs. While this is a functional feature for social media automation, it involves outgoing network requests to potentially untrusted remote sources.

x-twitter-automation