The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the belt-sh/cli and other skills from the inference-sh repository using the npx skills add command. It also links to external installation instructions hosted at https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md.
[COMMAND_EXECUTION]: The skill relies on the Bash(belt *) tool to run the belt CLI, which executes remote AI models and interacts with social media platforms (e.g., posting to Twitter).
[INDIRECT_PROMPT_INJECTION]: Multiple shell script examples in the documentation interpolate user-supplied or AI-generated variables ($i, $CONCEPT, $topic, $script) directly into shell command arguments without sanitization or escaping.
Ingestion points: Variables defined in bash for loops and script definitions within SKILL.md.
Boundary markers: None. The variables are placed directly inside string-interpolated JSON payloads passed to the CLI.
Capability inventory: The skill uses Bash to invoke the belt utility for remote model execution.
Sanitization: No escaping or input validation is present in the bash examples to prevent shell meta-character injection (e.g., if a variable value contained backticks or semicolons).

ai-social-media-content