korean-transit-route
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill uses the
npxutility to download and execute instructions and helper files from the vendor-owned package@nomadamas/k-skill@0. This is used to maintain up-to-date execution logic and retrieve necessary reference documents like legal disclaimers. - [COMMAND_EXECUTION]: The agent is instructed to run shell commands to source local environment variables and perform
curlrequests to the ODsay transit API. It also uses embedded Python snippets to securely encode API keys before transmission. - [INDIRECT_PROMPT_INJECTION]: The skill processes natural language location names from user input which are interpolated into external API calls.
- Ingestion points: User-specified departure and arrival addresses extracted from natural language prompts in
instruction.md. - Boundary markers: Not explicitly present in the interpolation logic for geocoding and routing URLs.
- Capability inventory: The skill performs network operations via
curland Python'surllib.request, and accesses local files via shell sourcing. - Sanitization: The skill correctly implements
urllib.parse.quote()to sanitize user-provided location strings before they are used in network requests, preventing URL-based injection attacks.
Audit Metadata