alterlab-bgpt-search
Warn
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's setup instructions specify using
npx mcp-remoteandnpx bgpt-mcp. These commands fetch packages from the public npm registry at runtime. - [REMOTE_CODE_EXECUTION]: Executing packages via
npx(specificallymcp-remoteandbgpt-mcp) from sources that are not verified trusted organizations allows for arbitrary code execution within the user's execution environment. - [COMMAND_EXECUTION]: The skill utilizes the
Bashtool and provides shell commands for the configuration and execution of the MCP server. - [PROMPT_INJECTION]: The skill processes data from a remote source (the BGPT MCP server), which could potentially contain malicious instructions designed to influence the agent's behavior (Indirect Prompt Injection). * Ingestion points: Full-text study data, methodology details, and conclusions retrieved from the
search_paperstool. * Boundary markers: No explicit markers or instructions are provided to the agent to distinguish or ignore potentially malicious content within the retrieved data. * Capability inventory: The skill has access to theBashtool, which increases the impact if an injection were to occur. * Sanitization: No sanitization, validation, or filtering of the remote content is specified.
Audit Metadata