The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from untrusted external sources including Meta, Google, and LinkedIn ad libraries, as well as organic social media content from TikTok, Instagram, and Reddit.
Ingestion points: Untrusted data enters the context through hermoso ads pull, hermoso research, and clone_video commands.
Capability inventory: The skill has access to the Bash tool, permitting the execution of local commands, network operations via the CLI, and publishing content to ten different social channels.
Boundary markers: No delimiters or protective instructions are specified to distinguish between processed data and agent instructions.
Sanitization: There is no mention of sanitization or validation of content retrieved from external ad libraries or social feeds.
[EXTERNAL_DOWNLOADS]: The skill uses npx -y hermoso to download and execute the vendor's tool from the npm registry. This provides the primary interface for all marketing tasks.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to run shell commands for research, asset generation, and campaign management. This includes potentially sensitive operations like auth login for credential management.
[DYNAMIC_EXECUTION]: The CLI enables dynamic tool discovery and execution. Commands like hermoso tools --search and hermoso call allow the agent to identify and run various vendor capabilities at runtime based on string matching.

hermoso-marketing