daangn-cars-search
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill utilizes the
npxpackage runner to fetch and execute the@nomadamas/k-skillCLI tool from the npm registry. This tool, owned by the skill author, is used to provide dynamic instructions, list helper files, and perform updates. - [COMMAND_EXECUTION]: The skill instructs the agent to run a bundled Python script (
scripts/daangn_cars.py) for car searches. The script relies entirely on the Python standard library and performs read-only data fetching from public web routes. - [INDIRECT_PROMPT_INJECTION]: The skill ingests car listing data (titles, descriptions, and region names) from the external site
daangn.com. While this creates an entry point for external data, the skill is scoped to read-only search operations and lacks the capabilities to perform sensitive actions based on that data without user oversight. - [SAFE]: The skill includes explicit instructions to never execute payments, messages, or transactions without immediate user approval. It also contains hard rules against handling plaintext credentials or bypassing security boundaries like CAPTHCAs.
Audit Metadata