The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes external content from Twitter/X (tweets, profiles, and search results), creating an attack surface for indirect prompt injection where malicious instructions could be embedded in social media data to influence the agent's behavior.
Ingestion points: Data is ingested from the Twitter/X platform via several endpoints in scripts/twitter_client.py and scripts/twitter_engagement_client.py (e.g., user info, timelines, search results).
Boundary markers: The skill lacks explicit boundary markers or instructions to the agent to disregard commands found within the retrieved text during processing.
Capability inventory: The skill has network access capabilities and the ability to read local workspace files for media uploads.
Sanitization: While the data is handled as structured JSON, there is no specific filtering or sanitization of the natural language content within that JSON to prevent command injection.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to an external API to facilitate its core functionality.
Evidence: The Python scripts communicate with api.aisa.one to fetch data and perform account actions. This domain is the vendor's official API endpoint as documented in the skill's metadata and README.

twitter-autopilot