naver-news-search

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses npx to download and run the author's CLI tool (@nomadamas/k-skill) to fetch the most recent instructions and disclaimer documents.\n- [COMMAND_EXECUTION]: The skill executes curl commands to access a news search API through the author's proxy server (k-skill-proxy.nomadamas.org).\n- [INDIRECT_PROMPT_INJECTION]: The skill processes news headlines and summaries from Naver. It addresses potential injection risks by instructing the agent to quote the source verbatim and by utilizing a proxy that strips HTML tags and entities. Evidence chain:\n
  • Ingestion points: The proxy endpoint https://k-skill-proxy.nomadamas.org/v1/naver-news/search provides external news data to the skill (instruction.md).\n
  • Boundary markers: None explicitly present in the data stream, but the instructions require citing the source directly.\n
  • Capability inventory: The agent can execute network reads and provide formatted output to the user (instruction.md).\n
  • Sanitization: The proxy server removes HTML tags and entities before the agent receives the data (instruction.md).\n- [DYNAMIC_EXECUTION]: The skill dynamically fetches its operational instructions at runtime using a package manager runner (npx).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 04:29 AM