* feat(tooling): enforce noUncheckedIndexedAccess from the base tsconfig
Completes #104600: the flag moves to tsconfig.json, the five per-lane
copies are removed, and test lanes carry one documented opt-out at their
shared parent (the ~4,400 fixture-indexing sites are a tracked
lane-by-lane follow-up). Any future lane inherits the flag by default.
Synthetic probes verify src rejects unchecked reads while test files
compile; all nine lane commands green.
* fix(memory-core): give the batch-call sort a defined element type
Type-aware lint analyzes test files under the base flag (its tsconfig
extends the root, not the test parent), so the widened call[0] made the
bare toSorted() fire require-array-sort-compare.
* feat(ui): auto-load older catalog transcript on scroll with a stable viewport
* fix(ui): surface the load-older fallback when auto-load is blocked on a short thread
* fix(ui): clamp oversized aggregated tool output in catalog transcripts
* fix(ui): stop catalog auto-pagination when an older read makes no progress
* fix(ui): continue catalog paging through advancing empty pages; exhaust only on a stalled cursor
* fix(ui): detect catalog cursor cycles to bound auto-pagination
* feat(gateway): propagate approvals to ancestor sessions with replay
Squash-rebased #103921 segment onto the native-clients tip on current main.
Session-scoped approval events publish sanitized pending/terminal
transitions to opted-in session audiences, with authoritative pending
replay on stream subscribe and durable-expiry reconciliation through the
owning manager. SessionApprovalEvent/Replay wire types export from the
approvals owner module; Swift models regenerated.
(cherry picked from commit 2d1dcf9747044710111d0c730fc46ba6013a165c)
(cherry picked from commit ccf88efd56b513d07599ffcee997bf0cddaf9adc)
(cherry picked from commit 8eb33f59f00ec9ee4625259b76bd7ae60d2cd480)
(cherry picked from commit c297cbc93c8401e9c79c20f242245a194f5dd236)
(cherry picked from commit 93d68f28601ff37d863da4db009504993a22533a)
(cherry picked from commit d285ccd8b2f212bb57e814107e7709c896e96b7d)
(cherry picked from commit 5c536982231402b1b58d1683794ea9948b32d077)
(cherry picked from commit 1646e5b6dc4dd20de54e10f110a5909be18d7d50)
(cherry picked from commit 35cf1b705b247fb282f4d56ffe74bd2fb238221f)
(cherry picked from commit 4f8ef3699de54d9135d0a36bee9c4c11f5a51173)
(cherry picked from commit 74fbdcc99d5fff062d67760a51dba4ee9e84479b)
(cherry picked from commit 222b285502681d2be14b1819798e9ee5555f431e)
(cherry picked from commit b9e744951f676b600e7c5322bfd04b0e99797c07)
(cherry picked from commit 9e904efde5b6407b2d4748313ffc3ea57b4f33a7)
(cherry picked from commit 678f2384fafa29ea3073aacf55de64e69f50565d)
(cherry picked from commit 72842e5cf6e9f4aba7ee8a68b48db0011e10db73)
(cherry picked from commit be74c25e80e84a1f065053766f23efb84822e811)
(cherry picked from commit 93ba8c4b09)
(cherry picked from commit 8f23761372)
* feat(gateway): fail-closed plugin and tool approval gates
Squash-rebased #103932 segment onto the ancestor-propagation tip on
current main. Plugin node.invoke approvals claim a one-shot allow-once
decision before handing execution authority to the policy, so observation
or retry cannot replay a consumed approval; sibling tool gates bind
approval ids to their originating reviewer identity.
(cherry picked from commit 122df0d75281f572c012b6484ed5daf085d1d577)
(cherry picked from commit f23d8ac240a8dcf2a42c4daaae962263975a0ae8)
(cherry picked from commit 8632fb6436a224dac7a9a9ef0216884530de2c24)
(cherry picked from commit d9fe2dd5c53665e5732f5f3da5ed1907a686ade8)
(cherry picked from commit bb139f2c8aa36ddad70413e1ef79ff491a6f2ecd)
(cherry picked from commit 9b3e056b68ea515c7d4baef269289b6670570480)
(cherry picked from commit b11e66b59a7af1f3b08712de06864aed64d349e4)
(cherry picked from commit a12916ee592d03dfa35e86a2aaede4476a5d4e5d)
(cherry picked from commit d699de840fc8346892b9ae244fe3c3e8c4413032)
(cherry picked from commit 9b7e5a9608f94732226776b801b3a2ac65329e05)
(cherry picked from commit 4b507593f1b2f10b4496984c84a2803b853e5c5e)
(cherry picked from commit 56408a186733c4eeb5aa76bd5907cde7b0cd3d5b)
(cherry picked from commit a4051d6d8353d39d6fa0b5d69c36e06c3cd3e796)
(cherry picked from commit 76829805a7135ed3a7984ae372edd0d872cb0dfc)
(cherry picked from commit a8b493f934b644ef8688a8044c2bdcdcc08d686d)
(cherry picked from commit aceb990597affb8388c341a10cb386ef96bed3e4)
(cherry picked from commit a29b0e75482470c53a9fb4fe3f204e14cf71868f)
(cherry picked from commit 2e3be08653)
(cherry picked from commit c9ae3d7202)
* fix: main-gate repairs for the durable-approvals stack
- android: capture createdAtMs on the pending exec-approval write at
registration; canonical readback after a refresh that already replaced
the visible rows was dropping the approval instead of surfacing the
still-pending reconciliation message (#104913 merged without this)
- android: generous CI timeout ceilings in GatewayExecApprovalRuntimeTest
- agents: hermetic model-discovery test via the plugins/provider-runtime
boundary (lazy plugin-runtime resolution hangs vitest workers since #104770)
- cli: usage-cost settle-budget test asserts the budget bound on every call
instead of pinning the poll count (fast hosts fit a second poll in 50ms)
- protocol coverage: allowlist session.approval for ios/android (native
review rides exec.approval push/nudge delivery)
- docs map, native i18n inventory, plugin-sdk api baseline regenerated
* fix(cron): keep isolated run status ok when delivery fails
A successful isolated cron agentTurn whose post-run delivery phase returns
an error collapsed the execution status into the delivery error, so the
outer scheduled run persisted status=error even though the isolated session
ended successfully.
Decouple execution status from delivery outcome in finalizeCronRun: when the
agent payload is non-fatal and the run was not aborted, a delivery dispatch
error now keeps status=ok and records the failure separately via
delivered/deliveryAttempted plus delivery diagnostics, which drive the
already-decoupled run-log deliveryStatus (not-delivered). Aborted runs and
fatal agent payloads keep their error status.
Fixes#94058
* fix(cron): keep deliberate delivery-target refusals as error status
The #94058 fix kept isolated turn status=ok when post-run delivery
fails, but the broad `status === "error"` check also caught the
#91613 keyless implicit last-target refusal, which is a deliberate
delivery-target guard (errorKind "delivery-target") and must stay
status=error. Exclude that errorKind so genuine delivery dispatch
failures still become ok while target-guard refusals remain errors.
* fix(cron): surface successful run delivery errors to run logs
A successful isolated agent turn keeps status=ok when post-run delivery
fails (#94058), but the ok/not-delivered resolveRunOutcome branch only
forwarded delivery diagnostics and dropped the delivery dispatch error.
That left lastDeliveryError empty and the finished-event deliveryError
field blank, so CLI/UI/API run logs could not show why delivery did not
land for an otherwise successful run.
Thread the delivery error on a dedicated CronRunOutcome.deliveryError
field from resolveRunOutcome through executeDetachedCronJob and
applyJobResult into resolveDeliveryState, so it persists as
lastDeliveryError and reaches the finished event (and thus the run log)
without conflating it with a run-level error (lastError stays empty on a
successful run).
Add a service/run-log readback test proving the delivery error survives
the cron boundary into a persisted run-log entry, and extend the
isolated-turn regression to assert the dedicated deliveryError field.
* fix(cron): preserve successful delivery outcomes
* fix(cron): preserve startup delivery errors
* fix(cron): retain best-effort delivery errors
* fix(cron): preserve delivery errors on fallback paths
* fix(cron): preserve delivery errors on fallback paths
---------
Co-authored-by: Alix-007 <li.long15@xydigit.com>
* feat(tooling): enforce noUncheckedIndexedAccess in the scripts lane
Burns down all 153 scripts-lane errors (bench aggregation, release
checks, i18n inventories, argv indexing) and flips the flag in
tsconfig.scripts.json. Direct-Node-executed release harness scripts use
local narrowing instead of workspace imports, which do not resolve
under plain node execution. Benchmark measured loops untouched.
* fix(scripts): import expect helpers via relative package sources
tsconfig path aliases resolve from cwd under tsx, so release wrapper
scripts running against old release target cwds could not resolve
@openclaw/normalization-core (not a linked root dependency). Relative
package-source imports match the established pattern on the adjacent
lines and are cwd-independent; old-target planning verified directly.
* feat(mac): swap dashboard titlebar buttons with sidebar state
The Control UI reports sidebar collapsed/width over a new openclawNav
WKScriptMessage; the titlebar accessory shows toggle+back/forward
right-aligned to the sidebar edge while expanded, and
toggle+search+new-session while collapsed. Search opens the command
palette and + opens the new-session surface via openclaw:native-*
events. Older gateway bundles that never report keep the shipped
toggle/back/forward layout.
Refs #105129
* docs: describe state-dependent mac titlebar buttons
* feat(nodes): route alerts by active computer
* fix(ci): allowlist node presence mobile coverage
* test(prompts): refresh node presence snapshots
* fix(macos): await presence reporter actor hop
* fix(macos): type presence test delivery state
* docs(nodes): add active presence guide
* docs(nodes): format presence troubleshooting
* improve(ui): compact the new-session screen on phones
The /new target controls condense on phones: Where and Worktree share one
line and the folder picker takes its own full-width line below, so long
paths stay readable instead of stacking three loose rows. The shared
welcome hero (start screen and /new) scales down on phones so the
composer and recent chats fit the first screenful, and the /new scroll
inset matches the chat thread's mobile padding.
* fix(ui): keep mobile target rows in DOM order
Review caught that order: 2 on the folder chip made visual and tab order
diverge on phones. The folder chip now flexes next to Where with a 140px
floor that pushes Worktree to the second line, so the compact layout
keeps visual, reading, and focus order aligned.
* fix(ui): wrap the worktree fields with their toggle
Groups the worktree toggle and its branch/name fields in one flex unit so
line breaks never separate the dependent fields from the toggle at any
width; review caught that the previous flex sizing let the branch control
wrap alone on wider phones.
* fix(tlon): cap SSE payload JSON.parse at 16 MiB to prevent OOM
* fix(tlon): cap SSE stream buffer and JSON.parse at 16 MiB to prevent OOM
* chore(tlon): add production-style SSE bounded proof script (#101274)
Adds extensions/tlon/proof-sse-bounded.mts which drives the real
UrbitSSEClient.processStream and processEvent against Node Readable streams
(not unit-test mocks) to demonstrate:
- normal SSE events are still delivered through the stream path;
- an unterminated stream that would grow beyond 16 MiB is rejected before
unbounded accumulation;
- a single SSE payload above 16 MiB is rejected before JSON.parse.
The script passes on this branch and fails 2/3 assertions when run against
origin/main's sse-client.ts, providing the before/after proof ClawSweeper
requested.
* chore(tlon): fix oxlint catch type in proof script (#101274)
* fix(tlon): reject oversized SSE chunk before buffer concatenation
Move the stream byte-limit check before buffer += chunkStr so a
single oversized chunk never lands in the pending buffer. The old
guard ran after concatenation, which still allowed the memory spike
this hardening is meant to prevent.
Add a single-oversized-chunk test to prove the guard fires before
the chunk is concatenated into the pending buffer.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* chore(tlon): remove PR-specific SSE bounded proof script
The proof script was review-only validation for #101274. Durable behavior
coverage lives in sse-client.test.ts (stream buffer bounding, oversized
chunk rejection before concatenation, 16 MiB boundary, normal delivery,
1000 small events). Per ClawSweeper P3 finding, drop the one-off script
from the plugin tree.
* fix(tlon): bound SSE event buffering safely
* test(tlon): avoid unsafe optional chaining
* fix(tlon): count split Unicode at SSE limit
* fix(tlon): parse split SSE delimiters at limit
---------
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Co-authored-by: Peter Steinberger <steipete@gmail.com>
Peter Steinberger