Red commit: `d8a6d42` ([CI
assertion](https://github.com/Kpa-clawbot/CoreScope/actions/runs/37695889461)).
Review-regression red: `f73643b` ([CI
assertion](https://github.com/Kpa-clawbot/CoreScope/actions/runs/37703995494)).
Fixes#2131. Partial fix for #1154: U2 only; other audit items remain
open.
Both node views explain distinct transmissions versus recorded
observations. The shared help is reachable by keyboard and touch, stays
readable at either scroll edge, and remains open when the pointer enters
its text. Escape dismisses help while preserving node context and focus;
a second Escape keeps the existing navigation behavior.
Count expressions and API requests are preserved. One controller per
nodes page handles one active help element and one queued animation
frame. Scroll/resize work is coalesced; listeners are removed on page
teardown. Work is constant relative to packet/node counts, with no
dataset scans or new settings/dependencies.
E2E assertion added: `tests/e2e/test-e2e-playwright.js:114`. Thirteen
cases cover exact counts, accessible description, keyboard/touch, hover
transfer, Escape, and viewport/scroll bounds. Existing metric assertions
are unchanged.
Browser verified: http://localhost:63508 using Chromium, controlled
fixtures and actual API data. Desktop/mobile screenshots retained
locally; additional checks exercised resizing and repeated page cleanup.
Local validation: 189 standalone suites; core browser 148 passed, 3
fixture-dependent skips (existing manual skips unchanged);
location/metric suite 6 passed; syntax, theme variables and lint passed
(0 errors, 92 existing warnings).
## Preflight overrides
- OpenClaw preflight/profile unavailable; repository checks and local
Chromium used.
- Windows ingestor symlink fixture lacks permission. Backend files are
unchanged; full Linux Go CI is required.
Red commit: `ac60248` ([two intended label
failures](https://github.com/Kpa-clawbot/CoreScope/actions/runs/37137429333)).
Original grouping red: `767da9b`
([CI](https://github.com/Kpa-clawbot/CoreScope/actions/runs/36353796690)).
Fixes#2073.
#2088 is merged. Rebased onto master `415362c`; all eight patches are
unchanged, including the test-first history.
Recent Adverts groups the bounded sample by authoritative `advert_kind`:
Flood, Mixed flood / direct (empty path), Direct (empty path), and Other
/ unknown. Missing evidence stays unknown. Each advert appears once,
preserving reception metadata, order, packet links, counts and
origin/history explanations. Wording describes observed remaining paths
without inferring original send mode or RF distance.
Validation at `6317baf`:
- All 185 standalone suites passed; syntax, XSS and lint passed (zero
errors, 92 existing warnings).
- Browser verified: `http://127.0.0.1:51827` (temporary Go fixture
server, now stopped), desktop pane/full view and mobile. Mixed evidence,
exact links and caveats checked again after pushing.
- The expanded fixture exposes the existing #2104 failure in the
unmodified core harness. Supplemental integration with #2105's exact
test correction passes 135 checks, with three existing skips. Both PRs
remain separate.
- [Fresh green
CI](https://github.com/Kpa-clawbot/CoreScope/actions/runs/37138366726)
passed; three independent automated reviews found no must-fix issues.
E2E assertion added: `tests/e2e/test-e2e-playwright.js:142`.
Grouping is O(n) over the existing 20-advert sample. No new requests,
settings, colors or dependencies.
## Preflight overrides
External OpenClaw runner/profile unavailable; repository checks and
actual local Chromium were used.
Red commit: `ca553e4` ([expected assertion
failure](https://github.com/Kpa-clawbot/CoreScope/actions/runs/37137426691)).
Fixes#2104.
The observation hex test now opens the existing grouped fixture
directly, selects A → B → A with stable-ID locators, and asserts the
selected ID and exact raw bytes after each render. Missing fixture data
fails explicitly. Distinct observation frames are seeded after schema
migration; this test's fixed sleeps and silent skips are removed.
Rebased onto master `415362c` after #2090 merged. Both patches are
unchanged. The previous #1122 layout blocker now passes against the
actual current server and assets.
Validation at `5db8017`:
- All 185 standalone frontend suites passed.
- Core browser: 133 passed, three existing skips; #1122 layout 6/6,
filter 11/11, grouped-collapse 4/4.
- Syntax, YAML, fixture SQL, XSS and lint passed (zero errors, 92
existing warnings).
- Fresh red CI reached the intended detached-observation assertion.
[Green
CI](https://github.com/Kpa-clawbot/CoreScope/actions/runs/37138187496)
passed; three independent automated reviews found no must-fix issues.
Only test code and fixture setup change; no configuration/customizer
implications.
## Preflight overrides
External OpenClaw runner/profile unavailable; repository checks and
actual local Chromium were used.
Red commits: `dc7dcbe`, `76e3217` ([assertion
failures](https://github.com/Kpa-clawbot/CoreScope/actions/runs/37059977385)).
Fixes#2041; supplies #2085's backend contract.
Records at most two route bits per retained advert. Node flood counts
now use the same evidence, including mixed and transport-flood adverts.
Analytics failures log/count errors while core observation, path and
liveness writes continue. The unrelated recent-advert limit clamp is
removed.
Direct adverts (empty path) describes the observed remaining path across
valid hash widths. Firmware can forward an advert down to an empty path;
this cannot establish original send mode or RF distance. The `zero_hop`
API spelling remains compatible.
Successful post-upgrade evidence writes survive observation replacement
and restart. Older overwritten frames and failed writes can leave gaps.
The UI explains automatic background backfill.
## Measured validation
Synthetic Windows workloads; timings are not production guarantees:
- 300K retained adverts, one-hour window: median 64.35→39.84 ms; 13.98
MB→145 KB allocated. Seven-day allocations unchanged; five paired
samples.
- 1M transmissions/11M observations: backfill 5m43s. Simple 10 Hz
writer: all 3,423 observations persisted; p99 28 ms, maximum 323 ms.
- Simulated pre-backfill cursor: retained evidence ready 4m21s; full
catch-up 6m41s, 122 cache invalidations. Normal completed-migration
startup bulk-loads current evidence.
Portable opt-in scale tests and window benchmarks are included. All 183
frontend suites, full server tests, targeted races, lint and real-API
desktop/mobile checks passed. Full [Linux
CI](https://github.com/Kpa-clawbot/CoreScope/actions/runs/37063210229)
passed at `36def9c`; three independent reviews are clear. Local Windows
testing hit the unchanged symlink-privilege test. No configuration
changes.
Preflight overrides: external script/OpenClaw profile unavailable;
repository checks and Chromium used.
Fixes#2086.
The #1468 WebSocket browser checks could fail when initial channel
loading completed between their separate before/after evaluations. An
orphan message added no channel, yet unrelated loading changed the count
from 0 to 5; the positive control could also lose its sentinel when
loading replaced the list.
Each check now captures before state, processes its packet, and captures
after state in one synchronous browser evaluation. All four original
assertions and both message payloads are unchanged. This updates one
test file only, with no production, dependency or configuration changes.
## Validation
- Delayed real API loading reproduces both failures on unchanged master.
- Both corrected checks pass with loading held and with loading
completed; each uses one evaluation.
- Parent independently ran the delayed-loading harness and full browser
suite: 131 passed, three existing fixture skips.
- All 183 standalone frontend suites passed. Syntax, inventory,
whitespace and privacy checks passed.
TDD justification: test synchronization repair only. Existing assertions
demonstrate the baseline failure; no production behavior or manufactured
failing test was added.
Browser checks used Chromium against unchanged production source.
OpenClaw profile/external preflight were unavailable; repository checks
ran directly, with a 60-second local navigation budget. Final CI must
pass before merge.
Moves 290 root test-*.js into tests/unit (177, listed in test-all.sh) and tests/e2e (113, classified in scripts/non-unit-tests.json), per #1981 and PR-D of #1385. Root goes from 348 entries to 48. test-all.sh and test-fixtures/ stay put. The inventory guard now fails if a test reappears in the root or sits in the wrong folder.
Verified independently of the diff: the invoked sets are unchanged (test-all.sh 177 before and after, deploy.yml 96 before and after, both identical as sets), and a full local run of test-all.sh on master and on the branch produced 4702 output lines each whose only differences are absolute paths, stack-trace line numbers shifted by the REPO_ROOT line, the inventory wording and two perf ratios. The guard was mutation-checked: a test back in the root, a unit suite in tests/e2e, and a suite dropped from test-all.sh each make it exit 1. CI run 35246304316 ran 97 suites from tests/e2e and is green.
Follow-up 9335c51d finished the instruction files: no bare root test command is left in AGENTS.md, the squad charters, .github or docs, and every tests/ path they name resolves.
Merged by the interim maintainer without a second human reviewer: CI and the local runs above are the independent checks.
Known and deliberately out of scope: 18 of the 113 files in tests/e2e are invoked by no runner at all, and one of them cannot run anywhere because it requires jsdom, which is not a declared dependency. Tracked separately.
n30nex