Audit-Verlauf

Most static detections are Markdown syntax, built-in module lists, or mapping examples and are false positives. Eight findings are confirmed because the skill directs agents to run package-manager commands with inferred package names. Default automatic installation and commit settings create supply-chain and repository-integrity risks.

Большинство статических срабатываний связано с синтаксисом Markdown, списками встроенных модулей или примерами сопоставления и является ложноположительным. Восемь находок подтверждены, поскольку навык предписывает агентам выполнять команды менеджеров пакетов с выведенными именами пакетов. Настройки автоматической установки и создания коммитов по умолчанию создают риски для цепочки поставок и целостности репозитория.

Most static alerts are false positives from Markdown code fences and package mapping tables. The real risk is the skill design: it installs inferred dependencies by default without prompting, which can introduce unreviewed third-party code. No prompt injection attempt was found in the reviewed files.

Static analysis reported many command, script, network, filesystem, and environment patterns, but most are markdown cookbook examples or exclusion lists rather than executable code. The confirmed risk is that the skill intentionally runs package-manager commands, contacts registries, and updates manifests, so it is publishable with a medium supply-chain warning.

Documentation-only skill providing guidance to AI assistants for dependency management. Contains only markdown files with code examples showing legitimate package manager commands. No executable code, scripts, network calls, credential access, or persistence mechanisms detected. All static findings are false positives triggered by markdown documentation patterns.

Documentation-only skill providing guidance to AI assistants for dependency management. Contains only markdown files with code examples showing legitimate package manager commands. No executable code, scripts, network calls, credential access, or persistence mechanisms detected. All static findings are false positives triggered by markdown documentation patterns.

Pure documentation-based skill with no executable code. Contains only markdown guidance for AI assistants on dependency management workflows. No scripts, network calls, credential access, or persistence mechanisms detected.

Pure documentation-based skill with no executable code. Contains only markdown guidance for AI assistants on dependency management workflows. No scripts, network calls, credential access, or persistence mechanisms detected.

Pure documentation-based skill with no executable code. Contains only markdown guidance for AI assistants on dependency management workflows. No scripts, network calls, credential access, or persistence mechanisms detected.