The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to download and execute code from the npm registry using the npx command for packages bgpt-mcp and mcp-remote. These are standard methods for deploying MCP servers.
[COMMAND_EXECUTION]: Setup instructions include shell commands for configuring the agent host, specifically using npx to launch the BGPT MCP server or connect to the remote SSE endpoint at https://bgpt.pro/mcp/sse.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process full-text experimental data from external scientific papers, which represents a potential surface for indirect prompt injection.
Ingestion points: Data is retrieved from the bgpt.pro API via the search_papers MCP tool and incorporated into the agent's context (SKILL.md).
Boundary markers: There are no explicit instructions to the agent to treat retrieved paper content as untrusted or to use delimiters to separate it from instructions.
Capability inventory: The skill allows the agent to execute searches and report findings; it does not explicitly grant file-write or general network capabilities beyond the MCP server's scoped tool.
Sanitization: No specific sanitization or filtering of the external study content is defined in the operating procedures.

bgpt-paper-search