CedarCopilot/cedar-mail

Branch: staging

test(eval): three instruction-eval harnesses and their negative results

closed#2714CedarCopilot

CedarCopilot wants to merge 2 commits into staging from chore/f8-eval-harnesses

[2026-09-17] merging staging to main#2818
5/5Verified as far as it ran
apps/server/.claude/skills/playbook-instruction-eval/aop-classification-eval.ts, apps/server/.claude/skills/playbook-instruction-eval/deal-close-task-cancel-eval.ts, apps/server/.claude/skills/playbook-instruction-eval/domain-return-shape-eval.ts, +2 more
feat(eval): add playbook-instruction-eval suite#2812
5/5Verified as far as it ran
apps/server/.claude/skills/playbook-instruction-eval/aop-classification-eval.ts, apps/server/.claude/skills/playbook-instruction-eval/deal-close-task-cancel-eval.ts, apps/server/.claude/skills/playbook-instruction-eval/domain-return-shape-eval.ts, +2 more
fix(f8): clear vendor-bucket recommendations, not just tasks#2713
Not scoredNot scored
docs/design/f8-agent-log.md
[2026-09-06] merging staging to main#2709
3/5Needs a look
docs/design/f8-agent-log.md
[2026-09-06] merging staging to main#2704
5/5Verified as far as it ran
docs/design/f8-agent-log.md
feat(f8): step 4 repair script for vendor buckets recorded as deals (NOT RUN)#2698
5/5Verified as far as it ran
docs/design/f8-agent-log.md
chore(f8): vendor-bucket task cleanup script + workstream design docs#2696
5/5Verified as far as it ran
docs/design/f8-agent-log.md

Libra has not linked any other changes to the same issue yet.

Landed before

Libra has not recorded a change landing before this one on the same base yet.

Landed after

Libra has not recorded a change landing after this one on the same base yet.