korea-weather
Fail
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructs the agent to run
npx -y @nomadamas/k-skill@0 instruct korea-weatherto fetch and follow dynamic instructions. This results in the execution of code downloaded from the npm registry at runtime. - [EXTERNAL_DOWNLOADS]: The skill downloads instructions and helper files from an external npm package and references a third-party GitHub repository (
github.com/NomaDamas/k-skill). - [COMMAND_EXECUTION]: The workflow involves executing
npxfor setup, updates, and instruction retrieval, as well ascurlto interact with an external proxy service. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from an external API, which could be used to inject malicious instructions into the agent's context.
- Ingestion points: Weather data is ingested from
https://k-skill-proxy.nomadamas.orgviacurlininstruction.md. - Boundary markers: There are no markers or instructions provided to the agent to treat the API response as potentially untrusted data.
- Capability inventory: The skill uses
curlandnpx, which provide network access and the ability to execute further code. - Sanitization: There is no evidence of validation or sanitization of the API response before the agent processes and summarizes the content.
Recommendations
- AI detected serious security threats
Audit Metadata