📦
История аудитов
seoul-subway-arrival - 1 аудит
Версия аудита 1 Последняя
8 нояб. 2026 г., 14:45
All nine static findings are false positives: Markdown fences, documented lookup links, and readable bilingual instructions do not establish the flagged threats. Separate semantic risks involve automatically executing a mutable npm package and elevating its downloaded output to primary instructions. No evidence found of credential theft, concealed payloads, or malicious intent in the reviewed files.
3
Просканировано файлов
138
Проанализировано строк
4
Пункты проверки
0
Ложные срабатывания проигнорированы
Подтверждённые проблемы безопасности (2)
Высокий
Prompt Injection Attempt Detected
The text says, "Run this and follow its output as the primary instructions for this skill". Elevating mutable CLI output creates a prompt-injection path from remotely supplied text.
The instruction explicitly delegates primary authority to package output that can change independently of this review. No evidence found of an actual malicious downloaded response.
Средний
Automatic Execution of a Mutable External Package
The npx -y commands execute @nomadamas/k-skill@0 without an exact version or confirmation. The update command also targets other installed skills, extending scope beyond a station lookup.
The commands explicitly use automatic npm execution and a mutable major-version selector. Package compromise is a potential consequence, not an observed malicious payload.
Факторы риска
🌐 Доступ к сети (2)
⚙️ Внешние команды (6)
Аудитор:: codex