What it does
Rules read like team conventions; output shows severity, confidence, and the method that caught each problem, worst first via perch issues.
Agent tooling
Perch
A semantic code linter that judges your source against configurable plain-language rules and prints per-finding confidence.
Bookmark: Perch
Keep this in your collection.
Leave a noteWhat would you try with this?
: Perch
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
How you can use it
Start by writing your project standards in simple sentences. For example, state that discounts cannot exceed 20 percent or that comments must match the program. A developer can then install the tool and save your plain-language sentences into a configuration file.
Your developer will need an API key, which is a private code from TypeSafe that lets software connect to its service. Running a scan sends the relevant code to TypeSafe for analysis. The tool then lists matching issues, ranked alongside a score showing how confident it is about each problem.