Marc-André Bombeck ca2064bfe9 test(coach): add two-layer Coach eval harness ...

A deterministic golden-set + red-team floor that gates every Coach change
offline, plus an opt-in live-judge layer that scores warmth and groundedness
with a pinned model only when COACH_EVAL_API_KEY is set.

- 42 golden cases across grounding, cross-metric, data-honesty, provider-parity
  and own-baseline taxonomies, graded by behaviour-anchored binary criteria.
- Structured-claim grounding helpers extend the numeric verifier from numbers
  to threshold / own-baseline / confident-verdict claims; the numeric API is
  unchanged so the existing grounding tests stay green.
- Red-team battery asserts the existing inbound refusal, outbound dose screen,
  acute red-flag escalation and learn-link guards fire.
- Nightly + manual workflow runs the deterministic floor always; the live judge
  step runs only when the secret is present and is non-blocking.

2026-06-27 10:14:52 +02:00

..

ISSUE_TEMPLATE

chore(release): v1.4.25

2026-05-14 23:49:32 +02:00

workflows

test(coach): add two-layer Coach eval harness

2026-06-27 10:14:52 +02:00

dependabot.yml

chore(github): expand dependabot to GitHub Actions + Docker

2026-05-10 15:23:17 +02:00

FUNDING.yml

docs: README presentation polish + community-file gaps

2026-06-05 03:01:36 +02:00

PULL_REQUEST_TEMPLATE.md

chore(github): add issue + PR templates

2026-05-10 15:23:15 +02:00