gpt-researcher

This skill is a developer-oriented reference for GPT Researcher, an LLM-based autonomous deep-research agent that conducts web and local research and produces cited reports using a planner-executor-publisher pattern. It solves the problem of understanding, extending, debugging, and integrating with the GPT Researcher codebase, serving as an in-repo guide for developers rather than an end-user research runner.

The SKILL.md provides a quick start (Python usage of the GPTResearcher class and commands to run the FastAPI backend and Next.js frontend), a key-file-location map, an architecture overview of the research flow, and core patterns including an eight-step feature-addition recipe and a template for adding new retrievers. It is backed by an extensive references library covering the REST and WebSocket APIs, core component signatures, architecture, prompts, adding features, MCP integration, flows, advanced patterns, deep research, configuration, multi-agents, and retrievers. The API reference documents endpoints for report generation, chat, report management, and file upload/output, along with the full GPTResearcher parameter set and MCP configuration examples (including passing a GITHUB_TOKEN via an environment variable to an MCP server).

The target users are developers integrating or extending GPT Researcher — adding retrievers, wiring MCP data sources, customizing research pipelines, or troubleshooting. Use cases include understanding the orchestrator, building a custom retriever, and configuring MCP-backed research. Credential handling shown is standard and safe: secrets like GITHUB_TOKEN are read from environment variables rather than hardcoded, and the documented behavior centers on read-oriented web research and local report generation.