The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill requires the user to run npx -y @nomadamas/k-skill@0 instruct rhwp-advanced to obtain instructions, which involves downloading and executing code from an untrusted NPM package at runtime.
[EXTERNAL_DOWNLOADS]: The skill relies on external components from untrusted sources, including the rhwp Rust crate installed via cargo install and pre-built binaries from an external GitHub repository (github.com/edwardkim/rhwp).
[PRIVILEGE_ESCALATION]: The skill utilizes the rhwp convert sub-command, which is explicitly documented as a tool to "unlock read-only HWPs," representing a bypass of document-level security controls.
[PERSISTENCE]: The skill provides an update mechanism via npx -y @nomadamas/k-skill@0 update, which facilitates persistent control over the local execution environment through remote code updates.
[INDIRECT_PROMPT_INJECTION]: The skill processes complex HWP/HWPX documents for debugging and structural analysis, creating a vulnerability surface where malicious instructions in document content could influence agent behavior.
Ingestion points: HWP/HWPX file paths passed to CLI tools in instruction.md.
Boundary markers: None identified in the prompt instructions.
Capability inventory: Shell command execution via rhwp, npx, and cargo.
Sanitization: No evidence of input sanitization or filtering for HWP document content.
[COMMAND_EXECUTION]: The skill uses various shell commands including mkdir for directory creation, open for viewing output, and the rhwp CLI for document manipulation.

rhwp-advanced