The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to download and run the @nomadamas/k-skill package from the NPM registry, which is used to manage instructions and execute the Python helper script.
[COMMAND_EXECUTION]: The instructions require the agent to execute shell commands using npx to fetch instructions, list files, and run the core spell-checking script (korean_spell_check.py).
[DATA_EXFILTRATION]: The Python script transmits user-provided text or the contents of local files to the external domain nara-speller.co.kr via HTTP POST requests to perform the spell check.
[INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it ingests and processes content returned from a third-party website. The script extracts correction suggestions and reasons from the HTML response and returns them to the agent context.
Ingestion points: The fetch_spell_check_html function in scripts/korean_spell_check.py retrieves content from nara-speller.co.kr.
Boundary markers: The skill does not implement specific delimiters or 'ignore' instructions for the data returned from the external service.
Capability inventory: The skill has network access (via urllib.request) and file read access (via pathlib).
Sanitization: The script uses strip_html to remove HTML tags from the service response, but does not sanitize the text content for potential instructions targeting the LLM.

korean-spell-check