feat: multi-turn eval harness with mocked MCP (all 4 skills) - #61

Merged

vahidlazio merged 8 commits into

mainfrom

feat/multi-turn-eval

Jul 31, 2026

Commits on Jul 30, 2026

feat: add multi-turn eval harness with mocked MCP

vahidlazio
and
claude
committed
26d6f76
fix: calibrate multi-turn scenario conversations

vahidlazio
and
claude
committed
e16506f
fix: strengthen consent gate in all 4 SKILL.md files

vahidlazio
and
claude
committed
846d5bf
feat: add multi-turn scenarios for all 4 skills + stronger assertions

vahidlazio
and
claude
committed
fa5d71d
fix: return null scores for non-applicable scorers

vahidlazio
and
claude
committed
3bedca5
fix: classification footer ordering + remove null scores

vahidlazio
and
claude
committed
16c2597
fix: add more conversation turns to Eppo happy-path

vahidlazio
and
claude
committed
130e781
fix: add buffer conversation turns to PostHog and Eppo scenarios

vahidlazio
and
claude
committed
1980326