From a16c984f9393db91426b43d8c5871911e4219bf3 Mon Sep 17 00:00:00 2001
From: aarroyo <beyondnet.peru@gmail.com>
Date: Tue, 4 Aug 2026 10:47:09 -0500
Subject: [PATCH] fix(robosoft): core-integration read the Tracker's gate shape
 off the Core
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Step 5 reported "no gates" against a Core that had evaluated six of them.
The Core was right; the assertion was wrong.

The canonical `EvaluationResult` nests per-kind results under `results`
(`gate`, `artifact`, `compliance`). It is the TRACKER's own DTO that
flattens them to `gates`, and this robot reaches the Core THROUGH the
gateway, so it receives the canonical shape and never the flattened one.

A second defect sat in the same helper and would have survived the first
fix: `summarizeVerdict` compared each verdict against `'failed'` while the
Core emits `FAIL`. Even given the right array it would have reported zero
failures on a run where all six gates failed — a green-looking summary
over a red result, which is worse than no summary at all.

Verified against a captured REST response rather than argued, and run
against the UNFIXED helper first so the change is known to matter:

    before  gates=0 · failed=0   (and the check said "no gates")
    after   gates=6 · failed=6

Checked for the same shape elsewhere in robosoft/: no other robot reads
`.gates` or compares against a lowercase verdict.

Found while running this robot against a live two-cluster stack for the
first time — it is excluded from the default runner and had never
executed. The same run found a real product defect (the MCP chart's
runAsUser, fixed in evolith#425); this one was the instrument.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---
 robosoft/robots/core-integration.robot.mjs | 26 ++++++++++++++++++----
 1 file changed, 22 insertions(+), 4 deletions(-)

diff --git a/robosoft/robots/core-integration.robot.mjs b/robosoft/robots/core-integration.robot.mjs
index 47904d34..2e59e34d 100644
--- a/robosoft/robots/core-integration.robot.mjs
+++ b/robosoft/robots/core-integration.robot.mjs
@@ -127,8 +127,18 @@ export default {
       // may be "failed" (the probe manifest has findings) — that is a REAL Core
       // evaluation, which is exactly what proves the round-trip.
       const data = unwrap(r.body);
-      check('Core returns a real evaluation verdict (gates evaluated)', r.body?.success === true && Array.isArray(data.gates), {
-        detail: Array.isArray(data.gates) ? summarizeVerdict(data) : r.body?.error?.message || 'no gates',
+      // `results.gate`, NOT `gates`. The Core's canonical EvaluationResult nests
+      // per-kind results under `results` (`gate`, `artifact`, `compliance`); it is
+      // the TRACKER's own DTO that flattens them to `gates`, and this robot talks
+      // to the Core through the gateway, so it gets the canonical shape.
+      //
+      // Asserting the flattened key made this step report "no gates" against a
+      // Core that had evaluated six of them — a defect in the instrument that read
+      // as a defect in the product. Measured against a captured REST response:
+      // `data.gates` is undefined, `data.results.gate` has 6 entries.
+      const gates = gatesOf(data);
+      check('Core returns a real evaluation verdict (gates evaluated)', r.body?.success === true && Array.isArray(gates), {
+        detail: Array.isArray(gates) ? summarizeVerdict(data) : r.body?.error?.message || 'no gates',
       });
     }
 
@@ -177,8 +187,16 @@ function parseMcpInner(body) {
 }
 
 /** One-line summary of an evaluation verdict from the Core's `data` payload. */
+/** The canonical EvaluationResult nests gates under `results.gate`. */
+function gatesOf(data) {
+  return Array.isArray(data?.results?.gate) ? data.results.gate : undefined;
+}
+
 function summarizeVerdict(data) {
-  const gates = Array.isArray(data.gates) ? data.gates : [];
-  const failed = gates.filter((g) => String(g.verdict).toLowerCase() === 'failed').length;
+  const gates = gatesOf(data) ?? [];
+  // The Core emits `FAIL`, not `failed`. Comparing against 'failed' reported
+  // zero failures on a run where all six gates had failed — a green-looking
+  // summary over a red result, which is worse than no summary.
+  const failed = gates.filter((g) => /^fail/i.test(String(g.verdict))).length;
   return `gates=${gates.length} · failed=${failed}`;
 }