twitter-automation

Pass

Audited by Gen Agent Trust Hub on Nov 18, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the installation of the belt CLI and references additional automation skills from external repositories (e.g., inference-sh/skills). These are required for the skill's primary functionality.\n- [COMMAND_EXECUTION]: The documentation provides multiple examples of executing the belt CLI tool to perform actions on Twitter/X, including a belt login command for authentication. The environment is constrained by the allowed-tools configuration to only permit belt commands.\n- [INDIRECT_PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection as it processes and transmits untrusted content (e.g., tweet text or direct messages).\n
  • Ingestion points: Content provided via the --input JSON flag in commands like x/post-create and x/dm-send.\n
  • Boundary markers: The use of JSON formatted input provides structural delimitation for input data.\n
  • Capability inventory: The skill enables the agent to post to public timelines, send messages, and follow/unfollow accounts on social media.\n
  • Sanitization: No explicit sanitization or filtering of the user-provided content is defined within the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 18, 2026, 08:26 PM