Repository navigation

feat(evaluations): add LD judge event support - #63

Merged

donei003 merged 18 commits into

mainfrom

feature/ld-judges-phase3

Sep 16, 2026

Merged

feat(evaluations): add LD judge event support#63
donei003 merged 18 commits into
mainfrom
feature/ld-judges-phase3

Commits on Sep 2, 2026
- feat(evaluations): add LD judge event support
  donei003
  committed
  ccfdc0a
- refactor(evaluations): criteria naming + shared judge scoring contract
  
  donei003
  and
  claude
  committed
  f63cb3c
- fix(evaluations): correct judge resolution, score validation, and rendering
  
  donei003
  and
  claude
  committed
  1eb1976
- refactor(evaluations): drop pydantic, run criteria concurrently
  
  donei003
  and
  claude
  committed
  46bfa00
- polish(evaluations): event logging, docstrings, judge scoring tests
  
  donei003
  and
  claude
  committed
  da69d33
Commits on Sep 9, 2026
- Capture and forward judge success direction (isInverted)
  
  donei003
  and
  claude
  committed
  6509971
- Compute judge verdict client-side instead of emitting successDirection
  
  donei003
  and
  claude
  committed
  b26a307
- Compute judge verdict client-side (isInverted) (#78)
  donei003
  authored
  da5a566
Commits on Sep 10, 2026
- Report criterion scores and let LaunchDarkly rule on them
  
  donei003
  and
  claude
  committed
  c49de2a
- Report criterion scores and let LaunchDarkly rule on them (#79)
  donei003
  authored
  a2cefe3
- fix(evaluations): dedup criteria case-insensitively, matching the API
  
  donei003
  and
  claude
  committed
  4126d7b
Commits on Sep 11, 2026
- TEMP(evaluations): hardcode judge successDirection for proxy testing
  
  donei003
  and
  claude
  committed
  16b5d22
- fix(evaluations): offline judge message_history must carry FORMATTING_INSTRUCTIONS
  
  donei003
  and
  claude
  committed
  0b81f4e
Commits on Sep 15, 2026
- fix(evaluations): route judge configs to a compatible handler
  
  donei003
  and
  claude
  committed
  1a5e099
- Revert "TEMP(evaluations): hardcode judge successDirection for proxy testing"
  
  donei003
  and
  claude
  committed
  462b712
- fix(evaluations): prefer an exact-provider judge handler over a wildcard
  
  donei003
  and
  claude
  committed
  c61b3a5
- docs(evaluations): document criteria, judge handlers, and scoring policy
  
  donei003
  and
  claude
  committed
  5f5a626
- docs(evaluations): use a customer-style judge key in examples
  
  donei003
  and
  claude
  committed
  ceca049

Commits

Commits on Sep 2, 2026

Commits on Sep 9, 2026

Commits on Sep 10, 2026

Commits on Sep 11, 2026

Commits on Sep 15, 2026