The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to download the ht-ml-pp-cli tool from the author's official GitHub and NPM repositories.
[REMOTE_CODE_EXECUTION]: Installation procedures utilize npx and go install to fetch and execute code from remote sources associated with the vendor (mvanhorn). These are documented steps for establishing the tool's environment.
[COMMAND_EXECUTION]: The skill is designed to execute the ht-ml-pp-cli binary to perform local file system operations (managing a SQLite database and configuration files) and network requests to the ht-ml.app service.
[DATA_EXFILTRATION]: The CLI includes a --deliver webhook:<url> feature that enables command output to be sent to external HTTP endpoints. While functional for data routing, this mechanism could be misused to transmit sensitive site keys or local data to external servers if not used with caution.
[PROMPT_INJECTION]: The skill ingests and parses local HTML files (via assets sync and scan commands), which creates an attack surface for indirect prompt injection. Maliciously crafted HTML content could attempt to influence agent behavior through the tool's output.
Ingestion points: Local HTML files parsed by the assets sync and scan subcommands.
Boundary markers: The instructions do not specify the use of delimiters or warnings to the agent when processing HTML data.
Capability inventory: The skill has access to shell execution (Bash), local file system access, and network capabilities.
Sanitization: No explicit sanitization or validation of the processed HTML content is described in the skill instructions.

pp-ht-ml