The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions direct the user to run npx -y @nomadamas/k-skill@0, which downloads a package from the npm registry to fetch instructions, update the skill, and execute the counting script. This introduces a dependency on an external package from a source that does not belong to a recognized trusted organization.
[REMOTE_CODE_EXECUTION]: The use of npx -y facilitates the immediate execution of remote code from an external registry. This pattern allows for the execution of arbitrary scripts provided by the downloaded package without prior local verification of the content.
[COMMAND_EXECUTION]: The script scripts/korean_character_count.js includes a --file argument that utilizes fs.readFileSync to read the contents of arbitrary files provided via command-line arguments. While the tool only outputs character statistics, the capability to read any file accessible to the process is a broad permission.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface that could be exploited for indirect prompt injection.
Ingestion points: The readInput function in scripts/korean_character_count.js reads text from command-line arguments, local files, or standard input.
Boundary markers: There are no markers or safety instructions implemented to differentiate between user data and potential instructions within the processed text.
Capability inventory: The skill can read local files and print to the console. It lacks network capabilities or dynamic execution of the input text.
Sanitization: No input sanitization is performed, as the script is designed to provide exact character and byte counts for the raw input.

korean-character-count