diff --git a/.github/scripts/replay-ui-summary-test.sh b/.github/scripts/replay-ui-summary-test.sh
new file mode 100755
index 0000000..b75e988
--- /dev/null
+++ b/.github/scripts/replay-ui-summary-test.sh
@@ -0,0 +1,125 @@
+#!/usr/bin/env bash
+# Drives replay-ui-summary.sh over the traces in testdata/ and checks the
+# summary it renders. Run under the flags GitHub Actions uses for a `run:`
+# block, because that is where a swallowed failure hides.
+#
+# testdata/replay-ui-real-run.jsonl is the first 10 steps of the dogfood run in
+# actions run 31873049857 on master, with the per-step `hierarchy` dumps and the
+# rowElements/tabElements extractors removed so the file stays readable. Nothing
+# else was touched. That run was green, and badgeCountMatchesThePanel judged
+# nothing in all 80 of its steps. The other two traces are written by hand.
+set -euo pipefail
+
+here="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+script="$here/replay-ui-summary.sh"
+testdata="$here/testdata"
+work="$(mktemp -d)"
+trap 'rm -rf "$work"' EXIT
+
+failed=0
+rendered=""
+status=0
+
+summarise() { # <case name> <output dir>
+  rendered="$work/$1.md"
+  : > "$rendered"
+  status=0
+  GITHUB_STEP_SUMMARY="$rendered" SEED=3 MAX_STEPS=80 \
+    bash -eo pipefail "$script" "$2" >/dev/null 2>"$work/$1.err" || status=$?
+}
+
+fail() {
+  echo "FAIL: $*" >&2
+  failed=1
+}
+
+expect_status() { # <want> <case>
+  [ "$status" = "$1" ] || fail "$2: exit $status, want $1"
+}
+
+expect_line() { # <line> <case>
+  grep -qxF -- "$1" "$rendered" || fail "$2: summary has no line '$1'"
+}
+
+expect_absent() { # <pattern> <case>
+  if grep -q -- "$1" "$rendered"; then fail "$2: summary should not mention '$1'"; fi
+}
+
+plant() { # <case name> <fixture> -> echoes the output dir
+  local dir="$work/$1/runs/20260815-075347"
+  mkdir -p "$dir"
+  cp "$testdata/$2" "$dir/trace.jsonl"
+  echo "$work/$1/runs"
+}
+
+# A green run of the real thing. Every count here was measured, not chosen.
+summarise real "$(plant real replay-ui-real-run.jsonl)"
+expect_status 0 real
+expect_line "- 10 steps recorded, 10 verified, 0 with violations" real
+expect_line "| selectedStepIsInRange | 10 | 0 |" real
+expect_line "| exactlyOneStepIsSelected | 10 | 0 |" real
+expect_line "| stepCountMatchesTheList | 10 | 0 |" real
+expect_line "| screenshotShowsTheSelectedStep | 10 | 0 |" real
+expect_line "| switchingTabsKeepsTheStep | 2 | 8 |" real
+expect_line "| badgeCountMatchesThePanel | **0** | 10 |" real
+expect_line "- \`badgeCountMatchesThePanel\` judged nothing on this run: no step had a violations badge on screen" real
+expect_absent "checked nothing" real
+
+# Every property judged at least once, including the one the real run never
+# reached. Without this the counts above are consistent with a guard that can
+# only ever return zero.
+summarise every "$(plant every replay-ui-every-property.jsonl)"
+expect_status 0 every
+expect_line "- 4 steps recorded, 3 verified, 0 with violations" every
+expect_line "| noUncaughtExceptions | 3 | 0 |" every
+expect_line "| screenshotShowsTheSelectedStep | 3 | 0 |" every
+expect_line "| switchingTabsKeepsTheStep | 2 | 1 |" every
+expect_line "| badgeCountMatchesThePanel | 1 | 2 |" every
+
+# The fuzzer sat on the run list: the step page never rendered, so nothing was
+# ever compared. This is the run that used to pass.
+summarise blind "$(plant blind replay-ui-nothing-rendered.jsonl)"
+expect_status 1 blind
+expect_line "| selectedStepIsInRange | **0** | 4 |" blind
+expect_line "| badgeCountMatchesThePanel | **0** | 4 |" blind
+expect_line "- \`exactlyOneStepIsSelected\` judged nothing on this run: no step had a row in the step list" blind
+grep -q "this run checked nothing" "$rendered" || fail "blind: no checked-nothing verdict"
+grep -q "the step page never rendered" "$work/blind.err" || fail "blind: nothing on stderr"
+
+# The run directory exists but the trace does not, and the glob matches nothing
+# at all. Both are the harness dying before it checked anything.
+mkdir -p "$work/empty-run/runs/20260815-075347"
+summarise empty-run "$work/empty-run/runs"
+expect_status 1 empty-run
+grep -q "no trace at" "$rendered" || fail "empty-run: no missing-trace line"
+
+summarise no-glob "$work/no-glob/runs"
+expect_status 1 no-glob
+grep -q "nothing under" "$rendered" || fail "no-glob: no missing-directory line"
+
+# A property this summary does not know about is a summary that silently counts
+# six of seven properties, which is the bug one level up.
+drifted="$(plant drift replay-ui-every-property.jsonl)"
+sed 's/badgeCountMatchesThePanel/badgeAgreesWithThePanel/g' \
+  "$testdata/replay-ui-every-property.jsonl" > "$drifted/20260815-075347/trace.jsonl"
+summarise drift "$drifted"
+expect_status 1 drift
+grep -q "these counts cannot be trusted" "$rendered" || fail "drift: no untrusted verdict"
+
+# Same for a reading the spec no longer declares: the count would quietly go to
+# zero and read as a UI that stopped rendering.
+sed 's/extract("violationBadges"/extract("violationCounters"/' \
+  "$here/../../replay-ui/sanderling/spec.ts" > "$work/renamed-spec.ts"
+rendered="$work/renamed.md"
+: > "$rendered"
+status=0
+GITHUB_STEP_SUMMARY="$rendered" SPEC="$work/renamed-spec.ts" \
+  bash -eo pipefail "$script" "$(plant renamed replay-ui-real-run.jsonl)" \
+  >/dev/null 2>&1 || status=$?
+expect_status 1 renamed
+grep -q "no longer declares violationBadges" "$rendered" || fail "renamed: no drift verdict"
+
+if [ "$failed" = 0 ]; then
+  echo "replay-ui-summary.sh: ok"
+fi
+exit "$failed"
diff --git a/.github/scripts/replay-ui-summary.sh b/.github/scripts/replay-ui-summary.sh
new file mode 100755
index 0000000..af0cf4c
--- /dev/null
+++ b/.github/scripts/replay-ui-summary.sh
@@ -0,0 +1,222 @@
+#!/usr/bin/env bash
+# Reads the replay-ui dogfood trace and reports, per property, how many steps
+# that property actually judged. Kept out of the workflow YAML so it can be run
+# by hand against a local run:
+#
+#   GITHUB_STEP_SUMMARY=/dev/stdout .github/scripts/replay-ui-summary.sh runs/dogfood
+#
+# `sanderling test` exiting 0 says only that no property returned false. Every
+# property in replay-ui/sanderling/spec.ts declines to judge when a reading it
+# needs is absent, which is right individually and useless in aggregate: a run
+# that never rendered the step page returns false nowhere and exits 0, so
+# checked-and-clean and checked-nothing arrive at the same green tick. Section 8
+# of docs/development/design-principles.md is the rule this leg was breaking.
+set -euo pipefail
+
+root="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
+output="${1:-runs/dogfood}"
+spec="${SPEC:-$root/replay-ui/sanderling/spec.ts}"
+summary="${GITHUB_STEP_SUMMARY:-/dev/null}"
+
+shopt -s nullglob
+run_dirs=("$output"/*/)
+shopt -u nullglob
+
+{
+  echo "### replay-ui dogfood"
+  echo
+  echo "- seed \`${SEED:-unset}\`, budget ${MAX_STEPS:-unset} steps"
+} >> "$summary"
+
+if [ ${#run_dirs[@]} -eq 0 ]; then
+  echo "- **nothing under \`$output/\`**: the run never started, so no property was ever evaluated" >> "$summary"
+  echo "replay-ui: no run directory under $output/, so there is nothing to judge" >&2
+  exit 1
+fi
+
+trace="${run_dirs[${#run_dirs[@]} - 1]}trace.jsonl"
+if [ ! -f "$trace" ]; then
+  echo "- **no trace at \`$trace\`**: the run wrote no steps, so no property was ever evaluated" >> "$summary"
+  echo "replay-ui: $trace does not exist, so there is nothing to judge" >&2
+  exit 1
+fi
+
+TRACE="$trace" SPEC="$spec" python3 - >> "$summary" <<'PY'
+import json
+import os
+import re
+import sys
+
+trace_path = os.environ["TRACE"]
+spec_path = os.environ["SPEC"]
+
+
+def toolbar(values):
+    reading = values.get("toolbar")
+    return reading if isinstance(reading, dict) else None
+
+
+def numbered(reading, key):
+    return reading is not None and reading.get(key) is not None
+
+
+def listing(values, name):
+    reading = values.get(name)
+    return reading if isinstance(reading, list) else []
+
+
+# One entry per property in replay-ui/sanderling/spec.ts, holding the guard that
+# property opens with. The trace records extractor values, not verdicts, so
+# counting the steps a property really judged means restating its guard here,
+# and that restatement is the risk this file carries: a guard that drifts from
+# its property would report evidence that does not exist. The two checks at the
+# bottom make the two drifts that CAN be caught loud rather than silent.
+#
+# noUncaughtExceptions carries no guard on purpose: it compares a count, not an
+# element, so it judges every step the verifier accepted.
+GUARDS = {
+    "noUncaughtExceptions": [],
+    "selectedStepIsInRange": [
+        ("a toolbar reporting a step and a step count",
+         lambda c, p: numbered(toolbar(c), "step") and numbered(toolbar(c), "stepCount")),
+    ],
+    "exactlyOneStepIsSelected": [
+        ("a row in the step list", lambda c, p: len(listing(c, "stepRows")) > 0),
+    ],
+    "stepCountMatchesTheList": [
+        ("a toolbar reporting a step count", lambda c, p: numbered(toolbar(c), "stepCount")),
+        ("a row in the step list", lambda c, p: len(listing(c, "stepRows")) > 0),
+    ],
+    "screenshotShowsTheSelectedStep": [
+        ("a toolbar reporting a step", lambda c, p: numbered(toolbar(c), "step")),
+        ("a screenshot in the before panel",
+         lambda c, p: c.get("beforeScreenshotStep") is not None),
+    ],
+    "switchingTabsKeepsTheStep": [
+        ("a tab selection that changed from the step before",
+         lambda c, p: p is not None and p.get("activeTabs") != c.get("activeTabs")),
+        ("a toolbar on both steps",
+         lambda c, p: p is not None and toolbar(p) is not None and toolbar(c) is not None),
+    ],
+    "badgeCountMatchesThePanel": [
+        ("a violations badge on screen",
+         lambda c, p: len(listing(c, "violationBadges")) > 0
+         and listing(c, "violationBadges")[0] is not None),
+        ("a violations panel on screen",
+         lambda c, p: len(listing(c, "violationPanelCounts")) > 0),
+    ],
+}
+
+# A zero here is the run failing to render, not the fuzzer getting unlucky:
+# every one of these needs only that the step page came up, which it does on the
+# first step of any working run. The other three are left to be reported. Two of
+# them are trajectory-dependent - switchingTabsKeepsTheStep needs a tab switch
+# between consecutive steps, badgeCountMatchesThePanel needs the fuzzer to land
+# on a violating step AND open the violations tab there - and a gate that
+# convicts on an unlucky seed reports a regression it has not found.
+MUST_RENDER = (
+    "selectedStepIsInRange",
+    "exactlyOneStepIsSelected",
+    "stepCountMatchesTheList",
+    "screenshotShowsTheSelectedStep",
+)
+
+steps = 0
+verified = 0
+violating = 0
+malformed = 0
+property_names = set()
+judged = {name: 0 for name in GUARDS}
+met_once = {name: [0] * len(conditions) for name, conditions in GUARDS.items()}
+
+# The trace records only the extractors that changed at a step, so an
+# extractor's value at any step is the last change recorded for it, starting
+# from null. Steps the verifier skipped advance nothing and are not evaluations.
+values = {}
+previous = None
+with open(trace_path, encoding="utf-8", errors="replace") as lines:
+    for line in lines:
+        if not line.strip():
+            continue
+        try:
+            step = json.loads(line)
+        except ValueError:
+            malformed += 1
+            continue
+        steps += 1
+        property_names |= set((step.get("residuals") or {}).keys())
+        if step.get("violations"):
+            violating += 1
+        if step.get("skipped_verification") or step.get("transitional"):
+            continue
+        for name, change in (step.get("extractor_changes") or {}).items():
+            values[name] = change.get("curr")
+        verified += 1
+        for name, conditions in GUARDS.items():
+            met = [condition(values, previous) for _, condition in conditions]
+            if all(met):
+                judged[name] += 1
+            for index, held in enumerate(met):
+                met_once[name][index] += 1 if held else 0
+        previous = dict(values)
+
+report = []
+report.append("- %d steps recorded, %d verified, %d with violations"
+              % (steps, verified, violating))
+if malformed:
+    report.append("- **%d unreadable line(s)** in %s" % (malformed, trace_path))
+report.append("")
+report.append("- judged: the property compared real values. declined: a reading it "
+              "needs was absent, so it returned true without checking anything.")
+report.append("")
+report.append("| property | judged | declined |")
+report.append("| --- | --- | --- |")
+for name in GUARDS:
+    count = judged[name]
+    report.append("| %s | %s | %d |"
+                  % (name, count if count else "**0**", verified - count))
+
+report.append("")
+for name, conditions in GUARDS.items():
+    if judged[name] or not verified:
+        continue
+    absent = [label for index, (label, _) in enumerate(conditions) if not met_once[name][index]]
+    report.append("- `%s` judged nothing on this run: no step had %s"
+                  % (name, ", nor ".join(absent) if absent else "what its guard needs"))
+
+blind = None
+if not verified:
+    blind = "%s records %d step(s) and not one of them was verified" % (trace_path, steps)
+else:
+    silent = [name for name in MUST_RENDER if not judged[name]]
+    if silent:
+        blind = ("%s declined on every step, so the step page never rendered and the "
+                 "exit code is not evidence about the replay UI" % ", ".join(silent))
+
+drift = []
+if property_names and property_names != set(GUARDS):
+    drift.append("the spec's properties are %s but this summary counts %s"
+                 % (", ".join(sorted(property_names)), ", ".join(sorted(GUARDS))))
+try:
+    with open(spec_path, encoding="utf-8") as handle:
+        declared = set(re.findall(r'extract\(\s*"([^"]+)"', handle.read()))
+except OSError as error:
+    drift.append("could not read %s to check its readings still exist: %s" % (spec_path, error))
+    declared = None
+if declared is not None:
+    gone = sorted({"toolbar", "stepRows", "beforeScreenshotStep", "activeTabs",
+                   "violationBadges", "violationPanelCounts"} - declared)
+    if gone:
+        drift.append("%s no longer declares %s, so the counts above describe readings "
+                     "that do not exist" % (spec_path, ", ".join(gone)))
+
+if blind:
+    report.append("- **this run checked nothing**: %s" % blind)
+for reason in drift:
+    report.append("- **these counts cannot be trusted**: %s" % reason)
+
+print("\n".join(report))
+for reason in ([blind] if blind else []) + drift:
+    print("replay-ui: %s" % reason, file=sys.stderr)
+sys.exit(1 if blind or drift else 0)
+PY
diff --git a/.github/scripts/testdata/replay-ui-every-property.jsonl b/.github/scripts/testdata/replay-ui-every-property.jsonl
new file mode 100644
index 0000000..145e5a4
--- /dev/null
+++ b/.github/scripts/testdata/replay-ui-every-property.jsonl
@@ -0,0 +1,4 @@
+{"extractor_changes":{"activeTabs":{"curr":"screenshot,screenshot","prev":null},"beforeScreenshotStep":{"curr":1,"prev":null},"extractor_0":{"curr":0,"prev":null},"extractor_1":{"curr":0,"prev":null},"stepRows":{"curr":[{"active":true,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false}],"prev":null},"toolbar":{"curr":{"step":1,"stepCount":5},"prev":null},"violationBadges":{"curr":[],"prev":null},"violationPanelCounts":{"curr":[],"prev":null}},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"op":"true"}},"step":1,"timestamp":"2026-08-15T07:53:47Z"}
+{"extractor_changes":{"activeTabs":{"curr":"violations,screenshot","prev":"screenshot,screenshot"},"beforeScreenshotStep":{"curr":4,"prev":1},"stepRows":{"curr":[{"active":false,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":true,"step":4,"violating":true},{"active":false,"step":5,"violating":false}],"prev":[{"active":true,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false}]},"toolbar":{"curr":{"step":4,"stepCount":5},"prev":{"step":1,"stepCount":5}},"violationBadges":{"curr":[1],"prev":[]},"violationPanelCounts":{"curr":[1],"prev":[]}},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"op":"true"}},"step":2,"timestamp":"2026-08-15T07:53:47Z"}
+{"extractor_changes":{"activeTabs":{"curr":"hierarchy,screenshot","prev":"violations,screenshot"},"violationBadges":{"curr":[],"prev":[1]},"violationPanelCounts":{"curr":[],"prev":[1]}},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"op":"true"}},"step":3,"timestamp":"2026-08-15T07:53:47Z"}
+{"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"op":"true"}},"skipped_verification":true,"step":4,"timestamp":"2026-08-15T07:53:47Z"}
diff --git a/.github/scripts/testdata/replay-ui-nothing-rendered.jsonl b/.github/scripts/testdata/replay-ui-nothing-rendered.jsonl
new file mode 100644
index 0000000..9099693
--- /dev/null
+++ b/.github/scripts/testdata/replay-ui-nothing-rendered.jsonl
@@ -0,0 +1,4 @@
+{"extractor_changes":{"activeTabs":{"curr":"","prev":null},"extractor_0":{"curr":0,"prev":null},"extractor_1":{"curr":0,"prev":null},"stepRows":{"curr":[],"prev":null},"violationBadges":{"curr":[],"prev":null},"violationPanelCounts":{"curr":[],"prev":null}},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"op":"true"}},"step":1,"timestamp":"2026-08-15T07:53:47Z"}
+{"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"op":"true"}},"step":2,"timestamp":"2026-08-15T07:53:47Z"}
+{"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"op":"true"}},"step":3,"timestamp":"2026-08-15T07:53:47Z"}
+{"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"op":"true"}},"step":4,"timestamp":"2026-08-15T07:53:47Z"}
diff --git a/.github/scripts/testdata/replay-ui-real-run.jsonl b/.github/scripts/testdata/replay-ui-real-run.jsonl
new file mode 100644
index 0000000..bf7cf71
--- /dev/null
+++ b/.github/scripts/testdata/replay-ui-real-run.jsonl
@@ -0,0 +1,10 @@
+{"extractor_changes":{"activeTabs":{"curr":"screenshot,screenshot","prev":null},"beforeScreenshotStep":{"curr":1,"prev":null},"extractor_0":{"curr":0,"prev":null},"extractor_1":{"curr":0,"prev":null},"stepRows":{"curr":[{"active":true,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false},{"active":false,"step":6,"violating":false},{"active":false,"step":7,"violating":false},{"active":false,"step":8,"violating":false},{"active":false,"step":9,"violating":false},{"active":false,"step":10,"violating":false},{"active":false,"step":11,"violating":false},{"active":false,"step":12,"violating":false},{"active":false,"step":13,"violating":false},{"active":false,"step":14,"violating":false},{"active":false,"step":15,"violating":false},{"active":false,"step":16,"violating":false},{"active":false,"step":17,"violating":false},{"active":false,"step":18,"violating":false},{"active":false,"step":19,"violating":false},{"active":false,"step":20,"violating":false},{"active":false,"step":21,"violating":false},{"active":false,"step":22,"violating":false},{"active":false,"step":23,"violating":false},{"active":false,"step":24,"violating":false},{"active":false,"step":25,"violating":false}],"prev":null},"toolbar":{"curr":{"step":1,"stepCount":25},"prev":null},"violationBadges":{"curr":[],"prev":null},"violationPanelCounts":{"curr":[],"prev":null}},"metrics":{"cpu_percent":0,"heap_bytes":2771019,"total_memory_bytes":3893507},"next_action":{"kind":"Tap","selector":"data-testid:tab","tap_point":{"x":774,"y":69},"x":774,"y":69},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/1","step":1,"timestamp":"2026-08-15T07:53:47.883914163Z"}
+{"extractor_changes":{"activeTabs":{"curr":"screenshot,hierarchy","prev":"screenshot,screenshot"}},"metrics":{"cpu_percent":0,"heap_bytes":3490520,"total_memory_bytes":6061916},"next_action":{"key":"right","kind":"PressKey"},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/1","step":2,"timestamp":"2026-08-15T07:53:48.356557191Z"}
+{"extractor_changes":{"activeTabs":{"curr":"screenshot,properties","prev":"screenshot,hierarchy"},"violationPanelCounts":{"curr":[0],"prev":[]}},"metrics":{"cpu_percent":0,"heap_bytes":3737998,"total_memory_bytes":6595962},"next_action":{"kind":"Tap","resolved_bounds":{"height":109,"width":31,"x":337,"y":308},"selector":"desc:select step 12","tap_point":{"x":352,"y":362},"x":353,"y":363},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/1","step":3,"timestamp":"2026-08-15T07:53:48.584625848Z"}
+{"extractor_changes":{"beforeScreenshotStep":{"curr":12,"prev":1},"stepRows":{"curr":[{"active":false,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false},{"active":false,"step":6,"violating":false},{"active":false,"step":7,"violating":false},{"active":false,"step":8,"violating":false},{"active":false,"step":9,"violating":false},{"active":false,"step":10,"violating":false},{"active":false,"step":11,"violating":false},{"active":true,"step":12,"violating":false},{"active":false,"step":13,"violating":false},{"active":false,"step":14,"violating":false},{"active":false,"step":15,"violating":false},{"active":false,"step":16,"violating":false},{"active":false,"step":17,"violating":false},{"active":false,"step":18,"violating":false},{"active":false,"step":19,"violating":false},{"active":false,"step":20,"violating":false},{"active":false,"step":21,"violating":false},{"active":false,"step":22,"violating":false},{"active":false,"step":23,"violating":false},{"active":false,"step":24,"violating":false},{"active":false,"step":25,"violating":false}],"prev":[{"active":true,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false},{"active":false,"step":6,"violating":false},{"active":false,"step":7,"violating":false},{"active":false,"step":8,"violating":false},{"active":false,"step":9,"violating":false},{"active":false,"step":10,"violating":false},{"active":false,"step":11,"violating":false},{"active":false,"step":12,"violating":false},{"active":false,"step":13,"violating":false},{"active":false,"step":14,"violating":false},{"active":false,"step":15,"violating":false},{"active":false,"step":16,"violating":false},{"active":false,"step":17,"violating":false},{"active":false,"step":18,"violating":false},{"active":false,"step":19,"violating":false},{"active":false,"step":20,"violating":false},{"active":false,"step":21,"violating":false},{"active":false,"step":22,"violating":false},{"active":false,"step":23,"violating":false},{"active":false,"step":24,"violating":false},{"active":false,"step":25,"violating":false}]},"toolbar":{"curr":{"step":12,"stepCount":25},"prev":{"step":1,"stepCount":25}}},"metrics":{"cpu_percent":0,"heap_bytes":3535304,"total_memory_bytes":7911280},"next_action":{"key":"right","kind":"PressKey"},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/12","step":4,"timestamp":"2026-08-15T07:53:48.820882198Z"}
+{"extractor_changes":{"beforeScreenshotStep":{"curr":13,"prev":12},"stepRows":{"curr":[{"active":false,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false},{"active":false,"step":6,"violating":false},{"active":false,"step":7,"violating":false},{"active":false,"step":8,"violating":false},{"active":false,"step":9,"violating":false},{"active":false,"step":10,"violating":false},{"active":false,"step":11,"violating":false},{"active":false,"step":12,"violating":false},{"active":true,"step":13,"violating":false},{"active":false,"step":14,"violating":false},{"active":false,"step":15,"violating":false},{"active":false,"step":16,"violating":false},{"active":false,"step":17,"violating":false},{"active":false,"step":18,"violating":false},{"active":false,"step":19,"violating":false},{"active":false,"step":20,"violating":false},{"active":false,"step":21,"violating":false},{"active":false,"step":22,"violating":false},{"active":false,"step":23,"violating":false},{"active":false,"step":24,"violating":false},{"active":false,"step":25,"violating":false}],"prev":[{"active":false,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false},{"active":false,"step":6,"violating":false},{"active":false,"step":7,"violating":false},{"active":false,"step":8,"violating":false},{"active":false,"step":9,"violating":false},{"active":false,"step":10,"violating":false},{"active":false,"step":11,"violating":false},{"active":true,"step":12,"violating":false},{"active":false,"step":13,"violating":false},{"active":false,"step":14,"violating":false},{"active":false,"step":15,"violating":false},{"active":false,"step":16,"violating":false},{"active":false,"step":17,"violating":false},{"active":false,"step":18,"violating":false},{"active":false,"step":19,"violating":false},{"active":false,"step":20,"violating":false},{"active":false,"step":21,"violating":false},{"active":false,"step":22,"violating":false},{"active":false,"step":23,"violating":false},{"active":false,"step":24,"violating":false},{"active":false,"step":25,"violating":false}]},"toolbar":{"curr":{"step":13,"stepCount":25},"prev":{"step":12,"stepCount":25}}},"metrics":{"cpu_percent":0,"heap_bytes":4913839,"total_memory_bytes":8552983},"next_action":{"duration_millis":250,"from_x":98,"from_y":149,"kind":"Swipe","to_x":98,"to_y":749},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/13","step":5,"timestamp":"2026-08-15T07:53:49.038789589Z"}
+{"metrics":{"cpu_percent":0,"heap_bytes":4615419,"total_memory_bytes":8970775},"next_action":{"key":"left","kind":"PressKey"},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/13","step":6,"timestamp":"2026-08-15T07:53:49.251047338Z"}
+{"extractor_changes":{"beforeScreenshotStep":{"curr":12,"prev":13},"stepRows":{"curr":[{"active":false,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false},{"active":false,"step":6,"violating":false},{"active":false,"step":7,"violating":false},{"active":false,"step":8,"violating":false},{"active":false,"step":9,"violating":false},{"active":false,"step":10,"violating":false},{"active":false,"step":11,"violating":false},{"active":true,"step":12,"violating":false},{"active":false,"step":13,"violating":false},{"active":false,"step":14,"violating":false},{"active":false,"step":15,"violating":false},{"active":false,"step":16,"violating":false},{"active":false,"step":17,"violating":false},{"active":false,"step":18,"violating":false},{"active":false,"step":19,"violating":false},{"active":false,"step":20,"violating":false},{"active":false,"step":21,"violating":false},{"active":false,"step":22,"violating":false},{"active":false,"step":23,"violating":false},{"active":false,"step":24,"violating":false},{"active":false,"step":25,"violating":false}],"prev":[{"active":false,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false},{"active":false,"step":6,"violating":false},{"active":false,"step":7,"violating":false},{"active":false,"step":8,"violating":false},{"active":false,"step":9,"violating":false},{"active":false,"step":10,"violating":false},{"active":false,"step":11,"violating":false},{"active":false,"step":12,"violating":false},{"active":true,"step":13,"violating":false},{"active":false,"step":14,"violating":false},{"active":false,"step":15,"violating":false},{"active":false,"step":16,"violating":false},{"active":false,"step":17,"violating":false},{"active":false,"step":18,"violating":false},{"active":false,"step":19,"violating":false},{"active":false,"step":20,"violating":false},{"active":false,"step":21,"violating":false},{"active":false,"step":22,"violating":false},{"active":false,"step":23,"violating":false},{"active":false,"step":24,"violating":false},{"active":false,"step":25,"violating":false}]},"toolbar":{"curr":{"step":12,"stepCount":25},"prev":{"step":13,"stepCount":25}}},"metrics":{"cpu_percent":0,"heap_bytes":3996016,"total_memory_bytes":8970784},"next_action":{"kind":"Tap","selector":"data-testid:tab","tap_point":{"x":670,"y":69},"x":670,"y":69},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/12","step":7,"timestamp":"2026-08-15T07:53:49.451285818Z"}
+{"metrics":{"cpu_percent":0,"heap_bytes":4830079,"total_memory_bytes":8974523},"next_action":{"kind":"Tap","selector":"data-testid:step-row","tap_point":{"x":170,"y":256},"x":170,"y":256},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/12","step":8,"timestamp":"2026-08-15T07:53:49.658417939Z"}
+{"metrics":{"cpu_percent":0,"heap_bytes":3772499,"total_memory_bytes":8974523},"next_action":{"kind":"Tap","selector":"data-testid:step-row","tap_point":{"x":170,"y":363},"x":170,"y":363},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/12","step":9,"timestamp":"2026-08-15T07:53:49.865911674Z"}
+{"extractor_changes":{"beforeScreenshotStep":{"curr":6,"prev":12},"stepRows":{"curr":[{"active":false,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false},{"active":true,"step":6,"violating":false},{"active":false,"step":7,"violating":false},{"active":false,"step":8,"violating":false},{"active":false,"step":9,"violating":false},{"active":false,"step":10,"violating":false},{"active":false,"step":11,"violating":false},{"active":false,"step":12,"violating":false},{"active":false,"step":13,"violating":false},{"active":false,"step":14,"violating":false},{"active":false,"step":15,"violating":false},{"active":false,"step":16,"violating":false},{"active":false,"step":17,"violating":false},{"active":false,"step":18,"violating":false},{"active":false,"step":19,"violating":false},{"active":false,"step":20,"violating":false},{"active":false,"step":21,"violating":false},{"active":false,"step":22,"violating":false},{"active":false,"step":23,"violating":false},{"active":false,"step":24,"violating":false},{"active":false,"step":25,"violating":false}],"prev":[{"active":false,"step":1,"violating":false},{"active":false,"step":2,"violating":false},{"active":false,"step":3,"violating":false},{"active":false,"step":4,"violating":true},{"active":false,"step":5,"violating":false},{"active":false,"step":6,"violating":false},{"active":false,"step":7,"violating":false},{"active":false,"step":8,"violating":false},{"active":false,"step":9,"violating":false},{"active":false,"step":10,"violating":false},{"active":false,"step":11,"violating":false},{"active":true,"step":12,"violating":false},{"active":false,"step":13,"violating":false},{"active":false,"step":14,"violating":false},{"active":false,"step":15,"violating":false},{"active":false,"step":16,"violating":false},{"active":false,"step":17,"violating":false},{"active":false,"step":18,"violating":false},{"active":false,"step":19,"violating":false},{"active":false,"step":20,"violating":false},{"active":false,"step":21,"violating":false},{"active":false,"step":22,"violating":false},{"active":false,"step":23,"violating":false},{"active":false,"step":24,"violating":false},{"active":false,"step":25,"violating":false}]},"toolbar":{"curr":{"step":6,"stepCount":25},"prev":{"step":12,"stepCount":25}}},"metrics":{"cpu_percent":0,"heap_bytes":4783792,"total_memory_bytes":8974552},"next_action":{"key":"right","kind":"PressKey"},"residuals":{"badgeCountMatchesThePanel":{"op":"true"},"exactlyOneStepIsSelected":{"op":"true"},"noUncaughtExceptions":{"op":"true"},"screenshotShowsTheSelectedStep":{"op":"true"},"selectedStepIsInRange":{"op":"true"},"stepCountMatchesTheList":{"op":"true"},"switchingTabsKeepsTheStep":{"name":"p6","op":"predicate"}},"screen":"/runs/20260815-075332/steps/6","step":10,"timestamp":"2026-08-15T07:53:50.075785262Z"}
diff --git a/.github/workflows/replay-ui.yml b/.github/workflows/replay-ui.yml
index f4b27fd..e72d477 100644
--- a/.github/workflows/replay-ui.yml
+++ b/.github/workflows/replay-ui.yml
@@ -121,20 +121,14 @@ jobs:
           SEED: ${{ inputs.seed }}
           MAX_STEPS: ${{ inputs.max-steps }}
 
+      # Exit 0 above means no property returned false. It does not mean any
+      # property was ever evaluated against real content: they all decline to
+      # judge when the elements they read are absent, so a run that never
+      # rendered the step page is green and worthless. This step is what tells
+      # the two apart, and it fails the job when nothing was judged.
       - name: Summarise
         if: always()
-        run: |
-          {
-            echo "### replay-ui dogfood"
-            echo
-            echo "- seed \`$SEED\`, budget $MAX_STEPS steps"
-            for dir in runs/dogfood/*/; do
-              [ -f "$dir/trace.jsonl" ] || continue
-              steps=$(wc -l < "$dir/trace.jsonl" | tr -d ' ')
-              violations=$(grep -c '"violations":\[' "$dir/trace.jsonl" || true)
-              echo "- $steps steps recorded, $violations step(s) with violations"
-            done
-          } >> "$GITHUB_STEP_SUMMARY"
+        run: .github/scripts/replay-ui-summary.sh runs/dogfood
         env:
           SEED: ${{ inputs.seed }}
           MAX_STEPS: ${{ inputs.max-steps }}
diff --git a/Makefile b/Makefile
index f86a2cf..50532e5 100644
--- a/Makefile
+++ b/Makefile
@@ -29,7 +29,7 @@ WEB_DIST := replay-ui/dist
 
 GOLINES := $(shell $(GO) env GOPATH)/bin/golines
 
-.PHONY: bootstrap proto sidecar sanderling sanderling-web sanderling-android sanderling-ios install test test-go test-browser test-companion test-kotlin test-spec-api spec-typecheck web-test web-typecheck web-build web-dev replay-dev docs clean release-cli release-npm-dry fmt fmt-go fmt-kotlin fmt-ts fmt-swift
+.PHONY: bootstrap proto sidecar sanderling sanderling-web sanderling-android sanderling-ios install test test-go test-browser test-companion test-kotlin test-spec-api test-ci-scripts spec-typecheck web-test web-typecheck web-build web-dev replay-dev docs clean release-cli release-npm-dry fmt fmt-go fmt-kotlin fmt-ts fmt-swift
 
 bootstrap:
 	$(GO) mod download
@@ -117,7 +117,7 @@ fmt-ts:
 fmt-swift:
 	xcrun swift-format format -i -r companion/Sources
 
-test: test-go test-kotlin spec-typecheck test-spec-api web-typecheck web-test
+test: test-go test-kotlin spec-typecheck test-spec-api web-typecheck web-test test-ci-scripts
 
 test-go:
 	$(GO) test $(GO_PACKAGES)
@@ -141,6 +141,11 @@ test-companion: $(COMPANION_EMBED) $(RUNNER_EMBED)
 test-kotlin:
 	ANDROID_HOME=$(ANDROID_HOME) $(GRADLE) :sidecar:test
 
+# The CI scripts that read a trace and decide whether a green leg is
+# evidence. bash and python3 only, which is all a runner has.
+test-ci-scripts:
+	.github/scripts/replay-ui-summary-test.sh
+
 test-spec-api:
 	cd pkg/spec && npm test --silent
 
diff --git a/docs/development/ci.md b/docs/development/ci.md
index db0e291..e82641c 100644
--- a/docs/development/ci.md
+++ b/docs/development/ci.md
@@ -136,6 +136,24 @@ hold for any trace and need no recalibrating when the fixture changes. The
 seventh is the stock `noUncaughtExceptions`, which asks nothing of the panels
 and only fails if the UI throws. Any violation fails the job.
 
+So does a run that judged nothing. Exit 0 says no property returned false, which
+is not the same as any property having been evaluated: each one declines to
+judge when the elements it reads are absent, so a run where the trace failed to
+serve, or where the fuzzer sat on the run list, renders nothing and passes.
+`.github/scripts/replay-ui-summary.sh` reads the trace and puts a per-property
+count of judged against declined steps in the job summary. Run it by hand with
+`GITHUB_STEP_SUMMARY=/dev/stdout .github/scripts/replay-ui-summary.sh runs/dogfood`.
+
+It fails the job when any of the four properties that need nothing beyond the
+step page having rendered - `selectedStepIsInRange`, `exactlyOneStepIsSelected`,
+`stepCountMatchesTheList`, `screenshotShowsTheSelectedStep` - judged nothing at
+all. The other three are reported and not gated, because a zero on them is a
+seed getting unlucky rather than a broken leg: `switchingTabsKeepsTheStep` needs
+a tab switch between consecutive steps, and `badgeCountMatchesThePanel` needs the
+fuzzer to land on a violating step and open the violations tab in that same
+step. On the first run measured this way (seed 3, 80 steps) that last one judged
+nothing at all, so the fixture reaches it far too rarely to be worth gating on.
+
 ## Reading a failure
 
 Both workflows upload their run directories as artifacts, and write the step