naming-house

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow (instruction.md) only ingests outsider-provided free text as explicit interview inputs—surname/birthdate/birthtime/gender/name candidates/preferences—then runs local tooling (naming-house) to score and interpret those user-supplied values.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's first required step is to run "npx -y @nomadamas/k-skill@0 instruct naming-house" (and a related "npx -y @nomadamas/k-skill@0 files naming-house"), which fetches and executes remote package code at runtime and delivers the primary instructions that control the agent's behavior.