what is github-high-potential fr?

GitHub High Potential is a tool that scans a wide range of sources across the internet every twelve hours to spot AI projects that seem to be gaining momentum, then writes up a summary report. It is designed to work both as a standalone command line tool and as a skill that the Codex AI assistant can trigger on its own. To build its report, the tool pulls in GitHub repository search results for keywords related to AI, large language models, agents, retrieval, inference, and machine learning generally, along with GitHub's own daily trending repository list. It also checks Hacker News for AI related discussions, recent AI focused papers from arXiv, an academic paper repository, and RSS feeds from AI company blogs and other tech news sources. If you provide a Product Hunt access token, it will pull in relevant posts from there as well. Running the included shell script handles the entire setup automatically: it finds a compatible Python interpreter, version 3.11 or newer, creates a virtual environment if one does not already exist, installs the project, and then generates a report saved as a dated markdown file inside a reports folder, split into a morning or afternoon version depending on when it runs. The window and depth of each report can be adjusted through optional environment variables, such as how many hours back to look and how many candidates to include, and a GitHub access token can be supplied to raise search rate limits. Beyond the quick start script, the project also installs as a command line tool with separate commands for collecting data, generating a report, or doing both at once. Each generated report includes when it was created, the time window it covers, how many candidates were found, a snapshot of current trends, ranked lists of notable projects, and a breakdown of which sources contributed the findings. Generated data files and reports are deliberately excluded from version control.