Skip to content

ClauJev 2.9.0

Latest

Choose a tag to compare

@Civalski Civalski released this 10 Nov 16:38
· 2 commits to main since this release

Changed

  • The command guard only asks before critical changes in production. In the week after 2.8.0 it
    asked 111 times, and nearly all were routine work already approved: pushing a feature branch, merging
    a PR, deleting a merged branch, running the deploy workflow, applying a schema migration, a commit or a
    build. Claude Code's own permissions already cover those. The fixed rule now covers only force push to
    main/master, npm publish, terraform destroy, kubectl delete, gh repo delete, aws s3 rb,
    aws rds delete-db-instance/cluster, supabase db reset --linked and DROP/TRUNCATE. Jev is asked
    whether the command is a critical, hard-to-undo change in production. Replayed on the 52 commands Jev
    flagged that week, 1 would still ask; 8 of 8 critical test cases ask.
  • The credentials question only flags leaks: a secret committed, posted or sent to a third party, or
    printed in full. Copying an .env between local folders or reading a token for a local test no longer
    asks (threshold 0.65 → 0.80).
  • The safety net used when Jev cannot answer follows the same line (rm -rf build, reset --hard and
    branch -D no longer ask; DELETE FROM, force push and rm -rf ~ still do).
  • The final check warns instead of blocking. It blocked 6 times that week and Claude changed something
    after only 1. Jev's judgement now shows as a message to you; only a failed verify command still blocks.
    stopBlocks: true brings the old behaviour back.
  • Rule review sees the user's request. A project rule like "no deploy or authentication in this phase"
    fired 4 times while the user was asking for exactly that. When the request explicitly asks for what a
    rule forbids, it is no longer flagged (p 0.92 → 0.07 in the replay; an unrelated request still flags, 0.89).