The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches real-time traffic data and CCTV metadata (such as location names and status descriptions) from external APIs at data.ex.co.kr and openapi.its.go.kr. If these upstream sources were compromised to include malicious instructions, the agent might attempt to follow them when summarizing the results for the user.
Ingestion points: The scripts/highway_traffic.py script fetches data from external JSON and XML APIs and prints it to stdout for the agent to read.
Boundary markers: Absent. The agent is instructed to summarize the results directly without explicit delimiters or warnings to ignore instructions within the data.
Capability inventory: The skill has the ability to execute shell commands via npx and read environment variables or local configuration files (~/.config/k-skill/secrets.env).
Sanitization: The script parses structured JSON and XML data, but it does not perform sanitization of the string content to prevent prompt injection.
[COMMAND_EXECUTION]: The skill relies on npx to execute its primary instruction-gathering and execution commands. This involves running the @nomadamas/k-skill CLI tool and a local Python helper script. This is standard functionality for this vendor's skills.
[EXTERNAL_DOWNLOADS]: The skill uses npx to fetch the @nomadamas/k-skill package from the NPM registry. It also fetches traffic data from the Korea Expressway Corporation and the National Transport Information Center. These are well-known/official services for the skill's stated purpose.

highway-traffic-status