Weekly GitHub repo health reports for your whole portfolio
A Crongent repo health routine reads activity across every repo in scope once a week and writes a digest: what shipped, what is stuck, which repos went quiet, and where docs like README or STATE files no longer match the code. Leaders get one report instead of fifty dashboards.
The problem
When you run many repos, nobody can see the whole portfolio. Problems surface late, in a client call.
How to set it up in Crongent
- Create a weekly routine and list the repos it covers.
- Pick a cheap, fast model since this is mostly reading.
- Send the digest to your webhook, or turn on writes and have it open an issue in a reports repo.
Example routine
- Name
- Portfolio digest
- Trigger
- Weekly, Friday 17:00
- Model
- google/gemini-3.5-flash
- Instructions
- Summarize the week per repo in two lines: shipped, blocked, risk. End with the three repos that need attention.