The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the agent to execute npx -y @nomadamas/k-skill@0 instruct market-kurly-search and explicitly tells the agent to follow the command's output as the 'primary instructions'. This allows for dynamic, unverified code and instruction execution at runtime from the npm registry.
[EXTERNAL_DOWNLOADS]: The skill requires the manual installation of the market-kurly-search npm package and uses npx to fetch additional helper files and updates from the @nomadamas scope, which is not a verified or trusted organization.
[PROMPT_INJECTION]: By directing the agent to disregard the local file and follow the output of a remote command as its primary logic, the skill creates a mechanism for instruction redirection that can bypass local safety checks or static analysis.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and parse data from api.kurly.com and the __NEXT_DATA__ script tags on www.kurly.com. This creates a vulnerability surface for indirect prompt injection.
Ingestion points: Data is fetched from api.kurly.com search endpoints and the www.kurly.com/goods/ detail pages (instruction.md).
Boundary markers: There are no explicit instructions or delimiters provided to the agent to treat the fetched product data as untrusted or to ignore instructions embedded within the search results.
Capability inventory: The skill uses the market-kurly-search package to perform network requests and returns results to the agent for processing.
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from the external web surfaces before it is interpolated into the agent's response context.

market-kurly-search