national-pension-workplace

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). instruction.md and scripts/national_pension_workplace.py show the skill calls k-skill-proxy’s GET /v1/national-pension/workplace using user-supplied free-text parameters (--name, optional --b-no) and the helper ingests the proxy’s response JSON at runtime, so outsider-authored text can be submitted via those CLI inputs before a specific item is selected.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's primary runtime step uses npx to fetch and execute the remote package "@nomadamas/k-skill@0" (e.g., "npx -y @nomadamas/k-skill@0 instruct national-pension-workplace"), which downloads and runs remote code that controls the skill's instructions.