Supercov
Code quality and test coverage for coding agents: Jev scores each source file so the agent knows what to fix first.
Code quality and test coverage for coding agents: Jev scores each source file so the agent knows what to fix first.
Supercov sends Jev questions about source-file properties and security checks. Its code validates the typed responses, then computes quality scores or reports flagged security surfaces; test coverage is measured separately.
Does this file contain at least one method long enough that a reader cannot hold its steps in mind at once, and that visibly does several separable things? A long method that is one flat sequence of trivial statements does not count.
One source file: its path and complete source text. The selected Noul checks whether it contains a method too long to follow as one unit that also does separable things.
Find eligible source files and prepare a request for each file or requested change.
Send Jev the catalog or security questions for the source being assessed.
Validate response question IDs and typed answer bounds; compute quality scores in code or retain security findings without averaging.
Run the project’s existing test command and turn uncovered paths into actionable follow-up queries.
The question results remain inspectable claims tied to individual files, and the application performs score composition and coverage analysis itself.