Skip to content

Patina v6.3.1

Choose a tag to compare

@devswha devswha released this 15 Jul 12:57
· 750 commits to main since this release

What's Changed

  • release: promote dev to main (audit fixes #602–#604 + repo hygiene) by @devswha in #608
  • docs(exit-codes): document exit 4 (persona safety gate) and 130 by @devswha in #609
  • research: pre-registered rewrite-efficacy study (Study 0 pilot) by @devswha in #610
  • fix(web): pre-spike hardening — KV TTL atomicity, follower cache-poll, client-IP trust order (#605–#607) by @devswha in #612
  • feat(pro-packs): licensed pack delivery — private-repo serving + patina pack CLI by @devswha in #611
  • feat(doctor): npm update check + document plugin auto-update toggle by @devswha in #613
  • research: rewrite-efficacy Study 1 — 4× language gap, structural tells confirmed, house-style fingerprint now significant by @devswha in #614
  • docs(examples): second pattern-25 success example across ko/en/ja/zh by @devswha in #615
  • docs(env): document LS lock-poll tunables (#612 follow-up) by @devswha in #616
  • release: promote dev to main — web hardening, pro pack delivery, rewrite-efficacy studies by @devswha in #617
  • feat(patterns): #34 False Agency across four languages (+ en #9 serial-negation) by @devswha in #618
  • docs(readme): lead with the study finding + gitignore symlink fix by @devswha in #619
  • release: promote dev to main — pattern #34, README study-led intro by @devswha in #620
  • research: Study 2 — structure pack fails its efficacy study; retired from manifest; over-editing guard ships by @devswha in #621
  • release: promote dev to main — Study 2 results, over-editing guard, pack retirement by @devswha in #622
  • fix(api): pro tier broken in production — claude-sonnet-5 rejects temperature by @devswha in #623
  • deploy: pro-tier temperature hotfix (rehearsal-found production bug) by @devswha in #624
  • feat(playground): Pro upsell CTA at free-quota exhaustion by @devswha in #625
  • deploy: quota-exhaustion Pro upsell (launch funnel) by @devswha in #626
  • research(prereg): Study 3 registration — structure-plan-step intervention by @devswha in #627
  • research(study3): plan→execute harness + analyzer + overnight supervisor by @devswha in #628
  • research(study3): raise supervisor pass cap 12→48 by @devswha in #629
  • research(prereg): judge-calibration side study registration by @devswha in #630
  • research(calibration): judge-calibration harness by @devswha in #631
  • research(calibration): judge-calibration results by @devswha in #632
  • research(design): panel v2 reconstitution (post-kimi) by @devswha in #633
  • research(study3): execute timeout 600s→900s by @devswha in #634
  • chore(gitignore): study3 + judge-calibration artifact dirs by @devswha in #635
  • research(study3): results — plan-step intervention rejected on all criteria by @devswha in #636
  • research(calibration): det document-verdict threshold (score>=35, research lane) by @devswha in #637
  • research: adopt deterministic panel v2 judge by @devswha in #639
  • Release 6.3.1 by @devswha in #640

Full Changelog: v6.2.0...v6.3.1