marketplace-mcp-install
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads project files and release binaries from the vendor's GitHub repository (
github.com/ilyautov/marketplaces-mcp-ru) viagit cloneorcurlto facilitate the installation. - [COMMAND_EXECUTION]: The skill executes local Python scripts (
install.py,serve.py) to configure MCP servers, create virtual environments, and manage application dependencies. - [PRIVILEGE_ESCALATION]: The instructions include removing macOS security attributes (
xattr -d com.apple.quarantine) for downloaded files, which bypasses system-level safety checks intended to protect against unverified software. - [INDIRECT_PROMPT_INJECTION]: The skill processes live data from external marketplace APIs (Wildberries, Ozon, Yandex Market, Avito), creating a potential surface for malicious instructions embedded in seller account data.
- Ingestion points: API responses from marketplace seller platforms (such as product details or order descriptions) retrieved in
SKILL.md. - Boundary markers: Absent; there are no instructions to the agent to delimit or ignore instructions that may be embedded in the fetched data.
- Capability inventory: The skill possesses file writing capabilities (
~/.marketplace-mcp/), subprocess execution (python3), and the ability to modify configuration files for various AI clients (e.g., Claude Desktop). - Sanitization: The instructions do not specify any sanitization, filtering, or validation steps for the data retrieved from external APIs.
Audit Metadata