The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill directs users to install the belt CLI from a remote repository and references an installation script hosted on GitHub (inference-sh/skills). These resources are associated with the primary service being automated.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user content, such as tweet text and direct messages, creating a surface for indirect prompt injection.
Ingestion points: Input fields for x/post-tweet, x/post-create, and x/dm-send in SKILL.md.
Boundary markers: Uses JSON structure for input formatting, but lacks specific instructions for the agent to ignore instructions embedded within the processed text.
Capability inventory: The skill utilizes the Bash tool to run belt CLI applications, which perform network operations and social media actions.
Sanitization: Input is passed through JSON-encoded strings which provides structural delimitation, though no semantic filtering of the content is performed.
[COMMAND_EXECUTION]: The skill is built around executing shell commands via the belt CLI to perform all its functions, as restricted by the allowed-tools frontmatter field.

twitter-automation