What this is

The shortest path to using Principia two ways: as a typed registry (consume @measurement/core reads) and as a measurement-engine API (/api/v1/*).

Prerequisites

• For registry reads: depend on @measurement/core and consume Principia's typed reads — do not re-implement the schema.
• For the API: an auth-gated Principia endpoint (per the portfolio service standard — REST + MCP). The exact base URL, key, and env var are in docs/specification/SPEC.md; the API is auth-gated, not public.

1. Score confidence (θ̂)

curl -X POST https://<base>/api/v1/methods/scoring/confidence \
  -H "Content-Type: application/json" \
  -d '{ "responses": [...], "model": { ... } }'
# → θ̂ (grid-MLE) + Fisher SE/CI + info_gain_next + threshold_met

2. Response diagnostics

curl -X POST https://<base>/api/v1/methods/response-diagnostics \
  -d '{ "responses": [...] }'
# → per-item INFIT/OUTFIT misfit flags + log-RT anomaly flags

3. Methods advisor & priors

# the /api/v1/* measurement engines: methods, priors, posterior,
# local-validity, assemble-battery, open-text-codings
curl https://<base>/api/v1/methods?construct=engagement

4. Consume the registry (typed)

Sibling apps read constructs/instruments/effect-sizes through @measurement/core-typed reads. If you discover an instrument Principia hasn't surveyed, file a Principia-side assignment to register it — don't silently expand the shared schema.

Honest scope

The engineering substrate is live (PRN-010/011/013); content coverage grows construct-family by construct-family (the engagement family is furthest along). Trust docs/research/PIPELINE_STATUS.md for what's populated, and SPEC.md for the exact API auth.

See also

Concepts · APIs & Data Contracts · Trust & Grading