fix(tests): clear NODE_TEST_CONTEXT for nested node --test in two suites (#1508)

test-foundation.sh and test-discord.sh ran a nested node --test that would
exit 0 on failure under a parent runner. Both clear the variable now, and
each has a check that fails the suite if it comes back. Darkwing wrote it,
Filbert approved it (e464be6c). Nine suites green on an index export.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

BUILD-LOG.md

@@ -3266,3 +3266,14 @@ reports 107/107 node tests and 22/22 mutations caught. The canonical
 suite list in AGENTS.md. Carried into A2 before genesis: N8, N11, N5, N10,
 N12 and Filbert's r1 notes P1 to P3 (agents/darkwing/work/queue-a2/).
 The nested `node --test` fix (N13) is a separate commit.
+
+## 2026-09-27: Nested node --test fix in two suites (N13, #1508)
+
+`test-foundation.sh` and `test-discord.sh` started a nested `node --test`
+that inherited NODE_TEST_CONTEXT, so under a parent runner a failing
+test there would exit 0. Both now clear it, and each has one new check
+that fails the suite if the clearing is lost (foundation 43 to 44,
+discord 63 to 64). Darkwing wrote it (n13.patch 00868b2f; the suites
+hash 60f04822 and 2ad3be74), and Filbert approved it (e464be6c). The
+nine suites passed on an index export. Latent until now: nothing ran
+those suites under a node runner.

agents/darkwing/work/queue-a1/n13-check.sh

@@ -0,0 +1,39 @@
+#!/usr/bin/env bash
+# N13 check (#1508): a failing nested `node --test` must fail test-foundation.sh
+# and test-discord.sh even when a parent runner's NODE_TEST_CONTEXT is set.
+# Runs in a scratch clone only: n13-check.sh CLONE (a clone of HEAD with
+# node_modules linked). It copies this checkout's two suites into the clone,
+# plants a failing test in each suite's test directory, and restores after.
+set -uo pipefail
+SRC="$(cd "$(dirname "$0")/../../../.." && pwd)"
+CLONE="${1:?usage: n13-check.sh CLONE}"
+[ "$(cd "$CLONE" && pwd)" != "$SRC" ] || { echo "refusing to run in the canonical checkout" >&2; exit 4; }
+cd "$CLONE" || exit 1
+
+PLANT='import { test } from "node:test";
+import assert from "node:assert/strict";
+test("N13 planted failure", () => assert.equal(1, 2));'
+FAILS=0
+expect() { # NAME WANT GOT
+    if [ "$2" = "$3" ]; then echo "ok    $1 (rc $3)"; else echo "FAIL  $1 (want rc $2, got $3)"; FAILS=$((FAILS+1)); fi
+}
+run() { # SUITE -> rc, output in /tmp/n13-<suite>-<tag>.txt
+    NODE_TEST_CONTEXT=child-v8 NO_COLOR=1 "scripts/test-$1.sh" >"/tmp/n13-$1-$2.txt" 2>&1
+}
+
+for pair in foundation:scripts/foundation discord:packages/discord/tests; do
+    suite=${pair%%:*} dir=${pair#*:}
+    git checkout -q -- "scripts/test-$suite.sh"
+    printf '%s\n' "$PLANT" >"$dir/zz-n13-planted.test.mjs"
+    run "$suite" head-planted; expect "$suite at HEAD, planted failure, parent context set" 0 $?
+    cp -p "$SRC/scripts/test-$suite.sh" "scripts/test-$suite.sh"
+    run "$suite" fixed-planted; expect "$suite fixed, planted failure, parent context set" 1 $?
+    rm -f "$dir/zz-n13-planted.test.mjs"
+    run "$suite" fixed-clean; expect "$suite fixed, no planted failure, parent context set" 0 $?
+    sed -i 's/node_tests() { env -u NODE_TEST_CONTEXT node --test/node_tests() { node --test/' "scripts/test-$suite.sh"
+    run "$suite" mutant-clean; expect "$suite with the clearing removed, no planted failure" 1 $?
+    git checkout -q -- "scripts/test-$suite.sh"
+done
+git status --short
+echo "n13 check: $FAILS failed"
+[ "$FAILS" -eq 0 ]

agents/darkwing/work/queue-a1/n13.md

@@ -0,0 +1,59 @@
+# N13: nested `node --test` in two suites (#1508)
+
+Darkwing, 2026-09-26. Filbert's note N13, put in DEFERRED by Sage as a
+small reviewed item before A2. Nothing is committed, staged or pushed.
+
+## The defect
+
+`scripts/test-foundation.sh:76` and `scripts/test-discord.sh:142` start a
+nested `node --test` without clearing `NODE_TEST_CONTEXT`. Under a parent
+test runner the nested run reports to that runner and exits 0 whatever its
+tests do. I reproduced it at HEAD 40a02d2b: with a failing test planted in
+each suite's test directory and `NODE_TEST_CONTEXT=child-v8` set, both
+suites exit 0. The check line reads `OK node --test ... (summary
+missing)`. In a control run of foundation without the variable, the same
+planted test fails the suite, so only a run under a parent runner is
+blind. I ran that control for foundation only.
+
+## The change
+
+`n13.patch` (sha256 00868b2f) changes only the two suites, +20 −2 lines.
+Each suite now has a `node_tests` function that runs `env -u
+NODE_TEST_CONTEXT node --test`, and uses it for its test run. Each also
+gets one new check. It writes a failing test into the sandbox, runs it
+through `node_tests` with `NODE_TEST_CONTEXT=child-v8` set, and requires
+exit 1 and `✖ planted failure` in the output. If someone drops the `env
+-u`, that check fails on every run, not only under a parent runner.
+
+| File | sha256 |
+|---|---|
+| `scripts/test-foundation.sh` | 60f04822 |
+| `scripts/test-discord.sh` | 2ad3be74 |
+| `n13-check.sh` | 6a231759 |
+
+## The check
+
+`n13-check.sh CLONE` runs in a scratch clone only and refuses the
+canonical checkout. For each suite it plants a failing test in the real
+test directory and runs the whole suite with `NODE_TEST_CONTEXT=child-v8`:
+
+| Suite | Case | Exit |
+|---|---|---|
+| foundation | HEAD, planted failure | 0 (the defect) |
+| foundation | fixed, planted failure | 1 |
+| foundation | fixed, no planted failure | 0 |
+| foundation | fixed but `env -u` removed, no planted failure | 1 |
+| discord | HEAD, planted failure | 0 (the defect) |
+| discord | fixed, planted failure | 1 |
+| discord | fixed, no planted failure | 0 |
+| discord | fixed but `env -u` removed, no planted failure | 1 |
+
+In the fixed runs with the planted failure, the suite prints `FAIL node
+--test ...` with the pass count and the planted test's name. The last row
+of each is the mutation: the new check alone fails the suite.
+
+All nine suites pass at 40a02d2b with this change and the queue A1
+candidate: foundation 44 and discord 64, one more check each than before.
+
+I found no other nested `node --test` in the suites. `test-queue.sh` and
+`queue-commit.sh` already clear the variable.

agents/darkwing/work/queue-a1/n13.patch

@@ -0,0 +1,44 @@
+diff --git a/scripts/test-discord.sh b/scripts/test-discord.sh
+index 6044100c..9ec4ddda 100755
+--- a/scripts/test-discord.sh
++++ b/scripts/test-discord.sh
+@@ -139,7 +139,16 @@ else
+ fi
+ 
+ # --- the seven offline groups ---
+-node --test --test-reporter=spec packages/discord/tests/ >"$SANDBOX/node-test.log" 2>&1
++# A nested `node --test` inherits a parent runner's NODE_TEST_CONTEXT, reports
++# to that runner and exits 0 whatever its tests do, so the suite clears it
++# (#1508 N13). The planted failing test proves a failure still fails here.
++node_tests() { env -u NODE_TEST_CONTEXT node --test "$@"; }
++mkdir -p "$SANDBOX/planted"
++printf '%s\n' 'import { test } from "node:test";' 'import assert from "node:assert/strict";' 'test("planted failure", () => assert.equal(1, 2));' >"$SANDBOX/planted/planted.test.mjs"
++NODE_TEST_CONTEXT=child-v8 node_tests "$SANDBOX/planted/" >"$SANDBOX/planted.log" 2>&1
++[ $? -eq 1 ] && grep -q '^✖ planted failure' "$SANDBOX/planted.log"
++check "a failing nested test fails the run under a parent runner's NODE_TEST_CONTEXT" $?
++node_tests --test-reporter=spec packages/discord/tests/ >"$SANDBOX/node-test.log" 2>&1
+ NODE_RC=$?
+ check "node --test packages/discord/tests/ ($(grep -E '^ℹ pass' "$SANDBOX/node-test.log" | tr -d '\n' || echo 'summary missing'))" $NODE_RC
+ if [ "$NODE_RC" -ne 0 ]; then
+diff --git a/scripts/test-foundation.sh b/scripts/test-foundation.sh
+index 251eb674..b94dab5c 100755
+--- a/scripts/test-foundation.sh
++++ b/scripts/test-foundation.sh
+@@ -73,7 +73,16 @@ done
+ check "checked-in demo bundles equal a fresh generation" $DEMO_OK
+ 
+ # --- unit, CLI, privacy, non-effect and fixture-index tests ---
+-node --test scripts/foundation/ >"$SANDBOX/node-test.log" 2>&1
++# A nested `node --test` inherits a parent runner's NODE_TEST_CONTEXT, reports
++# to that runner and exits 0 whatever its tests do, so the suite clears it
++# (#1508 N13). The planted failing test proves a failure still fails here.
++node_tests() { env -u NODE_TEST_CONTEXT node --test "$@"; }
++mkdir -p "$SANDBOX/planted"
++printf '%s\n' 'import { test } from "node:test";' 'import assert from "node:assert/strict";' 'test("planted failure", () => assert.equal(1, 2));' >"$SANDBOX/planted/planted.test.mjs"
++NODE_TEST_CONTEXT=child-v8 node_tests "$SANDBOX/planted/" >"$SANDBOX/planted.log" 2>&1
++[ $? -eq 1 ] && grep -q '^✖ planted failure' "$SANDBOX/planted.log"
++check "a failing nested test fails the run under a parent runner's NODE_TEST_CONTEXT" $?
++node_tests scripts/foundation/ >"$SANDBOX/node-test.log" 2>&1
+ NODE_RC=$?
+ check "node --test scripts/foundation/ ($(grep -E '^ℹ pass' "$SANDBOX/node-test.log" | tr -d '\n' || echo 'summary missing'))" $NODE_RC
+ [ "$NODE_RC" -ne 0 ] && grep -E "^✖|AssertionError" "$SANDBOX/node-test.log" | head -20

docs/SESSIONS.md

@@ -417,3 +417,4 @@ are never rewritten or removed; corrections are new entries.
 2026-09-26T23:58:21Z | Sage (T3 Claude Code, thread 1ef1e4f8) | #1507 CHAT-03 brief r2 | Rocko R2 revise (07b938fb), one blocking finding on stop proof vs interruption; V-1 accepted with limits, lead decision 25; r3 to Dewey. A2 carry-forward 773dbd75 accepted.
 2026-09-27T00:04:42Z | Filbert (T3 Claude Code, thread 9cb9731e) | Queue A1 (#1508) r1 re-review and N13 | approved: delta-r1 b733b894, manifest-r1 85a8a453, n13.patch 00868b2f; review agents/filbert/work/queue-a1-review-r1-2026-09-26.md e464be6c; three non-blocking notes
 2026-09-27T00:07:44Z | Sage (T3 Claude Code, thread 1ef1e4f8) | #1508 queue A1 commit | Filbert approved r1 (e464be6c); 20 files match 85a8a453; nine suites green on the index incl. queue; test-queue.sh added to AGENTS.md suites; lead decision 26 (P1-P3 to A2, N13-a declined). A2 started.
+2026-09-27T00:10:31Z | Sage (T3 Claude Code, thread 1ef1e4f8) | #1508 N13 commit | nested node --test fix in test-foundation/test-discord; pins 60f04822/2ad3be74 match; nine suites green on the index (foundation 44, discord 64); DEFERRED entry closed.

docs/plans/DEFERRED.md

@@ -163,17 +163,6 @@ at every gate. Started 2026-09-12 during the control board MVP.
   Sage asked the SetSpark lead to use a one-word role. The rule stays as
   it is. (2026-09-26, #1506)
 
-- **A nested `node --test` passes even when its tests fail.** A child
-  `node --test` inherits NODE_TEST_CONTEXT from its parent and exits 0
-  whatever its results. Darkwing found it building queue A1 and fixed that
-  path with `env -u NODE_TEST_CONTEXT`, plus a test. Other suites that
-  start `node --test` from inside a test run have not been checked, so
-  they may be hiding failures. (2026-09-26, #1508)
-  Filbert's A1 review (6933b885, N13) found two: `test-foundation.sh` line
-  76 and `test-discord.sh` line 142 start a nested `node --test` without
-  clearing NODE_TEST_CONTEXT. Latent today, because neither runs under a
-  parent runner. Darkwing fixes both after the A1 delta (lead decision 23).
-
 ## Queue
 
 Moved to `docs/plans/QUEUE.md` on 2026-09-13. This file holds only gaps.
@@ -187,4 +176,4 @@ Moved to `docs/plans/QUEUE.md` on 2026-09-13. This file holds only gaps.
 - Control board checked the bind address but not Host or Origin (2026-09-26, #1507, found by Dewey): closed by the board guard commit. `foreignRequest` refuses a non-loopback Host, a wrong port, userinfo or a path in Host, and any foreign Origin on every route. Dewey authored it, and Rocko approved it (`agents/rocko/work/board-guard-review-2026-09-26.md`, 5a12f08e). Rocko's one low finding: URL normalization accepts a root slash, empty userinfo and an empty query in Host, so the notes' wording overstates what the raw-syntax check refuses. No bypass was found.
 - Discord engine: leaked fake pi and the timeout gap in `busy` (2026-09-26, #1509): fixed in 3edb15eb (Darkwing 6b R2, Rocko approved, review ed5510a0). Tests stop the engine in `finally`. A turn pi never started now stops pi with exit 1 instead of guessing, and the unit restarts it. Known limit: see the open entry on wedge restarts.
 - Combined concurrent test hangs (2026-09-14, #1509): closed on evidence from 3edb15eb. At default concurrency the union passed 406/406 three times on the R2 snapshot and 408/408 on the commit's index, where clean HEAD had hung four runs out of four. The inherited-intermittency entry from 2026-09-15 stays open until a later run shows it is the same cause.
-
+- A nested `node --test` passes even when its tests fail (2026-09-26, #1508): closed. Queue A1 (34a72af9) clears NODE_TEST_CONTEXT on its own path. N13 fixes `test-foundation.sh` and `test-discord.sh`, each with a check that fails the suite if the clearing is lost. Darkwing wrote it (n13.patch 00868b2f) and Filbert approved it (e464be6c). Sage's grep on 2026-09-27 finds no other nested `node --test`: in scripts, only the two suites, `test-queue.sh` and `queue-commit.sh`, and in JS, only the queue test that covers the fix.

scripts/test-discord.sh

@@ -139,7 +139,16 @@ else
 fi
 
 # --- the seven offline groups ---
-node --test --test-reporter=spec packages/discord/tests/ >"$SANDBOX/node-test.log" 2>&1
+# A nested `node --test` inherits a parent runner's NODE_TEST_CONTEXT, reports
+# to that runner and exits 0 whatever its tests do, so the suite clears it
+# (#1508 N13). The planted failing test proves a failure still fails here.
+node_tests() { env -u NODE_TEST_CONTEXT node --test "$@"; }
+mkdir -p "$SANDBOX/planted"
+printf '%s\n' 'import { test } from "node:test";' 'import assert from "node:assert/strict";' 'test("planted failure", () => assert.equal(1, 2));' >"$SANDBOX/planted/planted.test.mjs"
+NODE_TEST_CONTEXT=child-v8 node_tests "$SANDBOX/planted/" >"$SANDBOX/planted.log" 2>&1
+[ $? -eq 1 ] && grep -q '^✖ planted failure' "$SANDBOX/planted.log"
+check "a failing nested test fails the run under a parent runner's NODE_TEST_CONTEXT" $?
+node_tests --test-reporter=spec packages/discord/tests/ >"$SANDBOX/node-test.log" 2>&1
 NODE_RC=$?
 check "node --test packages/discord/tests/ ($(grep -E '^ℹ pass' "$SANDBOX/node-test.log" | tr -d '\n' || echo 'summary missing'))" $NODE_RC
 if [ "$NODE_RC" -ne 0 ]; then

scripts/test-foundation.sh

@@ -73,7 +73,16 @@ done
 check "checked-in demo bundles equal a fresh generation" $DEMO_OK
 
 # --- unit, CLI, privacy, non-effect and fixture-index tests ---
-node --test scripts/foundation/ >"$SANDBOX/node-test.log" 2>&1
+# A nested `node --test` inherits a parent runner's NODE_TEST_CONTEXT, reports
+# to that runner and exits 0 whatever its tests do, so the suite clears it
+# (#1508 N13). The planted failing test proves a failure still fails here.
+node_tests() { env -u NODE_TEST_CONTEXT node --test "$@"; }
+mkdir -p "$SANDBOX/planted"
+printf '%s\n' 'import { test } from "node:test";' 'import assert from "node:assert/strict";' 'test("planted failure", () => assert.equal(1, 2));' >"$SANDBOX/planted/planted.test.mjs"
+NODE_TEST_CONTEXT=child-v8 node_tests "$SANDBOX/planted/" >"$SANDBOX/planted.log" 2>&1
+[ $? -eq 1 ] && grep -q '^✖ planted failure' "$SANDBOX/planted.log"
+check "a failing nested test fails the run under a parent runner's NODE_TEST_CONTEXT" $?
+node_tests scripts/foundation/ >"$SANDBOX/node-test.log" 2>&1
 NODE_RC=$?
 check "node --test scripts/foundation/ ($(grep -E '^ℹ pass' "$SANDBOX/node-test.log" | tr -d '\n' || echo 'summary missing'))" $NODE_RC
 [ "$NODE_RC" -ne 0 ] && grep -E "^✖|AssertionError" "$SANDBOX/node-test.log" | head -20