The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to download and execute instructions from the vendor's NPM package @nomadamas/k-skill. This is used to provide dynamic instructions and helper files relevant to the user's environment.
[COMMAND_EXECUTION]: Instructions direct the agent to execute shell commands (npx, curl) to fetch data and instructions. These commands are limited to the skill's specific purpose of querying air quality data.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied location strings (regionHint, stationName) which are used to query external APIs.
Ingestion points: Command-line arguments processed in scripts/fine_dust.py and query parameters defined in instruction.md.
Boundary markers: None specific, as inputs are handled as discrete parameter values for an API.
Capability inventory: Limited to network requests via urllib to specified government and vendor API domains; no shell execution or file system writes originate from these inputs.
Sanitization: Input values are URL-encoded before being used in network requests, preventing basic injection attacks.
[CREDENTIALS_SAFE]: The Python script handles API keys by retrieving them from environment variables (AIR_KOREA_OPEN_API_KEY) rather than hardcoding them. It also provides instructions for secure storage in a local environment file, adhering to standard security practices.

fine-dust-location