The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references external resources for installation and documentation, specifically pointing to the inference.sh GitHub repository (inference-sh/skills) and instructions hosted at raw.githubusercontent.com.
[COMMAND_EXECUTION]: The skill instructions heavily rely on executing shell commands via the belt CLI to interact with the inference.sh API apps (e.g., belt app run x/post-tweet).
[PROMPT_INJECTION]: The skill includes apps for retrieving external content, such as x/post-get and x/user-get, which ingest untrusted data from Twitter into the agent's context. This creates a surface for indirect prompt injection attacks.
Ingestion points: Retrieval of tweet content and user profile data via x/post-get and x/user-get.
Boundary markers: Absent. No specific delimiters or instructions are provided to the agent to treat retrieved text as untrusted data.
Capability inventory: The agent has access to the Bash tool with the belt CLI to perform network operations and social media actions.
Sanitization: Absent. There is no evidence of filtering or sanitization applied to the content retrieved from Twitter before it is processed by the agent.

twitter-automation