The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses shell script templates in the 'Multi-Platform Repurposing' and 'Batch Content Creation' sections that interpolate user-controlled variables directly into shell commands. This creates a vulnerability surface where malicious input could trigger unintended command execution.
Ingestion points: The $CONCEPT and $TOPICS variables are defined within the skill context and interpolated into shell commands.
Boundary markers: No delimiters or escaping mechanisms are used to separate user data from the command structure.
Capability inventory: The skill utilizes shell execution via the belt CLI tool, which is scoped in the allowed-tools frontmatter.
Sanitization: There is no evidence of input validation, escaping, or sanitization for the interpolated shell variables.
[EXTERNAL_DOWNLOADS]: The skill references external resources and installation scripts.
Evidence: Instructions point to https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md for installation and suggest adding the belt-sh/cli tool via npx.
[COMMAND_EXECUTION]: The skill provides numerous pre-formatted shell commands for content generation using the belt CLI tool. While intended for legitimate functionality, these commands are executed in the local shell environment.

ai-social-media-content