What is hermes-researcher-agent?

This repository packages a researcher profile for Hermes Agent, an AI assistant platform created by Nous Research. The profile is designed for people who need an AI system that can search public sources broadly, check the quality of what it finds, and return a practical recommendation rather than a raw list of links. It is built around what the author calls an evidence gate: before giving a serious answer, the agent checks whether each source is credible, current, and free of important caveats. The profile is restricted to public sources by default. It does not scrape pages that require a login, use cookies, or access private data. This makes it safe to share across teams or machines without worrying that someone else's private sessions or credentials are bundled inside. The README is explicit about what this is not: it is not a private intelligence kit or a scheduled monitoring service out of the box. Practical use cases include comparing tools or vendors, checking whether a GitHub repository has real user adoption beyond star counts, scanning official documentation and changelogs, and producing decision-ready briefs that summarize findings with clear verdicts and stated limitations. The repository includes two Python helper scripts: one checks the public metadata of any GitHub repository (stars, forks, latest release, license), and another runs a diagnostic to verify which public data sources the agent can actually reach before a research session starts. Installation works through the Hermes command-line interface using a one-line command that pulls the profile from GitHub. After installing, the user connects their own AI model provider and selects which tool categories to enable, such as web search, browser access, or file saving. The profile itself does not supply any model or API key. Example research briefs generated with this profile are included in the repository, covering questions like whether the repository itself is ready to promote publicly and what the upstream Hermes Agent project looks like in terms of adoption. Documentation is provided in both English and Russian. The license is MIT.