W14: env-parity ledger — runner-only bindings vs fresh sx_server (test-only)

Section-B audit, all verified live over the epoch protocol. Runner-only
bindings absent from production: values, call-with-values (run_tests.ml
:1131/:1140), contains-char? (rt.ml:728 + rt.js:85), trim-right (JS runner
ONLY — absent even from the OCaml runner), sha3-256 (rt.ml:745 + rt.js:88;
production's real primitive is crypto-sha3-256).

Consequences pinned: (canonical-serialize 42) on a fresh server errors
"Undefined symbol: contains-char?" — content addressing broken for ANY
number outside the runners. And BOTH runners' sha3-256 are FAKE stubs
(OCaml: Hashtbl.hash), so every test-computed CID differs from production.

scripts/test-env-parity.sh is a bidirectional ledger: MUST_HAVE bindings
going missing fail; a KNOWN_DRIFT binding APPEARING also fails with
instructions to move it to MUST_HAVE and flip the consequence pin — the
ledger cannot rot silently in either direction. 7/7 green.

Test-only: no semantics edits, no push.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

plans/agent-briefings/sx-gate-loop.md

@@ -59,9 +59,11 @@ Pin each confirmed-and-fixed finding with a minimal repro. Add suites to
       conformance — "housekeeping" was a mislabel from F-15's tag)
 
 ### B. Runner/production env unification
-- [ ] Audit runner-only bindings (`values`/`call-with-values` F7/K42, JS
+- [x] Audit runner-only bindings — inventory + bidirectional ledger in
-      fake sha3/equal?/apply/env-set! shims JS5) — inventory + failing pin
+      `scripts/test-env-parity.sh` (KNOWN_DRIFT: values, call-with-values,
-      that a fresh `sx_server` reproduces the drift
+      contains-char?, trim-right, sha3-256; consequence pin:
+      canonical-serialize broken on server; BOTH runners' sha3-256 are FAKE
+      stubs → test CIDs ≠ production CIDs)
 
 ### C. Harness honesty
 - [ ] K19 — MCP `mcp_tree.ml` harness primitive table drift vs `sx_primitives`
@@ -83,6 +85,20 @@ Pin each confirmed-and-fixed finding with a minimal repro. Add suites to
 
 ## Progress log (newest first)
 
+- 2026-07-04 — **Section B: env-parity audit + ledger**. Probed a fresh
+  `sx_server` over the epoch protocol (`deps-check` + live eval). Confirmed
+  runner-only drift: `values`/`call-with-values` (run_tests.ml:1131/1140),
+  `contains-char?` (rt.ml:728 + rt.js:85), `trim-right` (**JS runner only**
+  — absent even from the OCaml runner), `sha3-256` (rt.ml:745 + rt.js:88).
+  Consequence verified live: `(canonical-serialize 42)` on the server →
+  `Undefined symbol: contains-char?` (content addressing broken for ANY
+  number outside the runners). **Worse than the finding**: BOTH runners'
+  `sha3-256` are FAKE stubs (OCaml uses `Hashtbl.hash`!) while production
+  has real `crypto-sha3-256` — every CID computed in tests differs from
+  production CIDs. Authored `scripts/test-env-parity.sh` as a bidirectional
+  ledger: MUST_HAVE regressions fail; a KNOWN_DRIFT binding *appearing*
+  also fails (forces ledger + consequence-pin update when W5/W7/W12 land
+  fixes). 7/7 green. Test-only.
 - 2026-07-04 — **S4 error-page-cache pin (item A.7) — section A COMPLETE**.
   Extended `scripts/test-protocol-gate.sh` with an HTTP-mode case: fresh
   `sx_server.exe --http <random-port>` (timeout-bounded, own PID killed at

scripts/test-env-parity.sh

@@ -0,0 +1,100 @@
+#!/bin/bash
+# test-env-parity.sh — W14 section-B ledger: runner env vs production env.
+#
+# The review (F7, K42, JS5, core.md "canonical.sx depends on test-runner-only
+# helpers") found bindings that exist ONLY in the test runners, so suites
+# pass against an environment production never provides. Rule (PLAN.md W14):
+# "if the spec needs it, it's a kernel primitive; if not, the test can't
+# have it."
+#
+# This script is a LEDGER, not a wish: it asserts today's confirmed drift
+# stays exactly as recorded. Both directions fail loudly:
+#   - a MUST_HAVE going missing on the server  -> regression, fix the kernel
+#   - a KNOWN_DRIFT binding appearing on the server -> the fix landed;
+#     move it to MUST_HAVE and update the consequence pins below.
+#
+# Confirmed inventory (2026-07-04, all verified live over the epoch protocol):
+#
+#   binding            OCaml runner        JS runner    fresh sx_server
+#   values             real (rt.ml:1131)   ?            ABSENT
+#   call-with-values   real (rt.ml:1140)   ?            ABSENT
+#   contains-char?     real (rt.ml:728)    real (:85)   ABSENT
+#   trim-right         ABSENT              real (:87)   ABSENT
+#   sha3-256           FAKE Hashtbl.hash   FAKE stub    ABSENT (real = crypto-sha3-256)
+#
+# Consequences (pinned in section 3):
+#   - (canonical-serialize 42) on a fresh server errors "Undefined symbol:
+#     contains-char?" -> content addressing broken for ANY number outside
+#     the test runners.
+#   - every CID computed inside run_tests uses a FAKE hash, so test CIDs
+#     never equal production CIDs (crypto-sha3-256 is real SHA3).
+#
+# Each probe spawns its OWN timeout-bounded sx_server.exe. No shared process.
+set -uo pipefail
+
+cd "$(dirname "$0")/.."
+SERVER=hosts/ocaml/_build/default/bin/sx_server.exe
+
+if [[ ! -x "$SERVER" ]]; then
+  echo "SKIP: $SERVER not built (run sx_build target=ocaml first)" >&2
+  exit 2
+fi
+
+pass=0
+fail=0
+
+# deps_unresolved EXPR -> prints the (unresolved ...) list for EXPR on a fresh server
+deps_unresolved() {
+  printf '(epoch 1)\n(deps-check "%s")\n' "$1" \
+    | timeout 60 "$SERVER" 2>/dev/null \
+    | grep -o ':unresolved ([^)]*)' || true
+}
+
+# --- Section 1: MUST_HAVE — spec-needed bindings production must provide ---
+MUST_HAVE_EXPR='(list (equal? 1 1) (apply + (list 1 2)) (contains? {:a 1} :a) (crypto-sha3-256 \"x\") (split \"a-b\" \"-\"))'
+unres=$(deps_unresolved "$MUST_HAVE_EXPR")
+if [[ -z "$unres" || "$unres" == ':unresolved ()' ]]; then
+  echo "PASS: MUST_HAVE core bindings all resolve on fresh sx_server"
+  pass=$((pass+1))
+else
+  echo "FAIL: MUST_HAVE binding missing on fresh sx_server: $unres"
+  fail=$((fail+1))
+fi
+
+# --- Section 2: KNOWN_DRIFT — runner-only bindings, asserted ABSENT -------
+# If one of these starts resolving, its kernel fix landed: move it to
+# MUST_HAVE above and update the consequence pin in section 3.
+for name in values call-with-values contains-char? trim-right sha3-256; do
+  unres=$(deps_unresolved "($name)")
+  if grep -q -- "$name" <<<"$unres"; then
+    echo "PASS: KNOWN_DRIFT '$name' still absent on fresh sx_server (ledger accurate)"
+    pass=$((pass+1))
+  else
+    echo "FAIL: KNOWN_DRIFT '$name' now RESOLVES on fresh sx_server — fix landed?"
+    echo "      Update this ledger: move '$name' to MUST_HAVE and revisit section 3."
+    fail=$((fail+1))
+  fi
+done
+
+# --- Section 3: consequence pin — canonical.sx on the production server ---
+# Current reality: canonical-serialize of ANY number errors on a fresh
+# server because canonical-number calls runner-only contains-char?.
+out=$(printf '(epoch 1)\n(load "spec/canonical.sx")\n(epoch 2)\n(eval "(canonical-serialize 42)")\n' \
+  | timeout 60 "$SERVER" 2>&1)
+if grep -q 'error 2 .*contains-char?' <<<"$out"; then
+  echo "PASS: consequence pin — canonical-serialize on numbers still broken on server (as recorded)"
+  pass=$((pass+1))
+elif grep -q '^(ok 2 ' <<<"$out"; then
+  echo "FAIL: consequence pin — canonical-serialize 42 now WORKS on the server."
+  echo "      The canonical-helpers fix landed: flip this pin to assert success"
+  echo "      and pin the exact canonical form + CID stability."
+  fail=$((fail+1))
+else
+  echo "FAIL: consequence pin — unexpected server output:"
+  sed 's/^/  /' <<<"$out"
+  fail=$((fail+1))
+fi
+
+echo
+echo "env-parity: $pass passed, $fail failed"
+[[ $fail -eq 0 ]]