---
name: adapter-impl-question-triage
description: "BINDING DRI protocol (operator 2026-07-24): an adapter builder's impl question is ALWAYS docs-under-read / docs-gap / core-parity-gap — identify with evidence, fix the PUBLIC contract, NEVER hand them spt-core source."
metadata: 
  node_type: memory
  type: feedback
  originSessionId: b5bfa053-0922-4c48-9e08-fe64cebcde7e
  modified: 2026-07-24T08:43:26.857Z
---

**Operator-directed 2026-07-24, added to doyle /role.** Adapters — omp-spt ([[emphasys-agent]]), claude-spt (perri) — build INDEPENDENTLY of spt-core source, blind from the PUBLIC docs at http://localhost:5474.

**When an adapter builder asks an implementation question, it is ALWAYS exactly one of three:**
1. **Not reading the public docs closely enough** — the docs already answer it. (Point them at the section.)
2. **DOCS GAP** — the public contract is incomplete, so every adapter implements it blind and wrong. (Amend the public docs.)
3. **Worst case: spt-core MISSING a feature** the adapter needs for legacy-spt parity. (Orchestrate a core amendment.)

**Why:** an adapter that must read core source to build correctly is a boundary failure — [[spt-core-harness-boundary-and-grounding]] / [[adapter-glue-model-boundary]] / [[consumer-verifies-core-decides]]. The public contract must be self-sufficient.

**How to apply:**
- Identify WHICH of the three with SOLID EVIDENCE: read the public docs (`docs-site/src/`) + CONTEXT.md. NEVER hand the builder spt-core source internals.
- Answer ONLY from the public contract. If incomplete → AMEND THE PUBLIC DOCS (durable fix for ALL adapters), add/tag the traceable REQ's `doc` stage, run the docs-drift gate (`cargo run -p xtask -- check`), THEN point them at the amended section.
- WITHHOLD implementation details adapters must not depend on (internal thresholds, nonce derivations); steer them to implementation-independent approaches (e.g. synthetic-input regression tests, not hardcoded magic numbers).
- CHECK EVERY DECISION against CONTEXT.md — the design source of truth, especially the harness-agnostic boundary: **spt-core never caters to a harness.**

**First application (the pattern that minted this):** emphasys asked 7 EVENT-PART reassembly questions; I nearly answered from `chunk.rs` source. Correct: it was a DOCS GAP (public docs named the tag but omitted every reassembly rule → omp-spt wedged implementing blind). Fixed by amending messaging/overview.md + REQ doc-stage; answered from the contract; withheld the 400 threshold + hex-nonce id as impl. See [[daemon-lifecycle-progress]] EVENT-PART resolution entry.