Навыки youtube-transcript История аудитов
📦

История аудитов

youtube-transcript - 1 аудит

Версия аудита 1 Последняя

14 нояб. 2026 г., 11:24

All 286 static alerts match documented authentication, consent-based setup, ordinary shell operations, or Markdown; no evidence found of malicious intent. The system-prompt alert quotes explicit defensive guidance, not an extraction request. Two additional credential-handling risks concern existing file targets and unredacted signup responses after parsing failures.

21
Просканировано файлов
909
Проанализировано строк
6
Пункты проверки
0
Ложные срабатывания проигнорированы

Подтверждённые проблемы безопасности (2)

Средний
Existing Key Targets Are Not Secured Before Overwrite
The key write uses umask 077 but does not secure existing files or reject symbolic links. Existing targets retain permissions; a link to a readable external target can expose the saved key despite the private parent directory.
Shell redirection follows symbolic links, and umask affects creation rather than existing file modes. Exposure requires a preexisting unsafe target; newly created ordinary files receive private permissions.
Средний
Signup Parsing Failure Can Disclose Credentials
A format-specific sed expression extracts the signup key. If extraction fails, the script prints the full response, which may still contain a valid credential in another JSON representation.
The failure branch prints RESPONSE without redaction after a regex rejects it. Disclosure depends on the response representation; no live credential disclosure was observed.

Факторы риска

⚙️ Внешние команды (50)
🔑 Переменные окружения (50)
🌐 Доступ к сети (13)
📁 Доступ к файловой системе (21)
Аудитор:: codex