* improve(pr): trust origin/main-matching wrappers when the canonical checkout is parked elsewhere
* improve(crabbox): hint at lease expiry when reused-lease runs fail fast
* fix(pr): pin the wrapper trust anchor to the remote-tracking ref
* feat(tooling): extend strict-ratchet lane to all remaining leaf packages
* fix(packages): burn down indexed-access debt in ai and agent-core
All 196 noUncheckedIndexedAccess errors fixed behavior-identically
(iteration, .at()/slices, DataView byte math, guarded queue peeks) plus
cleanup finds: a sparse-tool diagnostic bug, deduplicated Responses
call/item ID parsing with an honest string|undefined splitter, and a
clearer local failure for redacted-thinking blocks missing signatures.
Removes all 17 remaining non-null assertions across ai, gateway-client,
and llm-core so the assertion ban aligns 1:1 with the ratchet lane.
speech-core joins memory-host-sdk as structurally deferred (imports
src/** via plugin-sdk path mappings).
* feat(tooling): add noUncheckedIndexedAccess strict-ratchet lane
* fix(packages): make indexed access explicit across ratchet packages
Burns down all 65 noUncheckedIndexedAccess errors in the seven ratchet
packages with behavior-identical restructuring (iteration, charAt/slice,
regex-group guards, validated hextet tuple). net-policy invariant
violations now throw instead of failing open. Adds
@openclaw/normalization-core/expect (expectDefined/first/last) with
subpath export and tests. memory-host-sdk fixes kept but the package
stays out of the lane: it re-exports core src/** so the flag would apply
transitively to all of core.
* fix(lint): keep .oxlintrc.json strict-JSON parseable for extension lint wrappers
* feat: node-hosted plugins — dynamic tools, MCP servers, and skills
Nodes become declarative plugin hosts:
- node.pluginTools.update: node hosts publish plugin-registered agent tool
descriptors; gateway materializes them as agent tools executing via
node.invoke under the node command allowlist, with tools.effective
invalidation and node online/offline removal.
- Trusted paired-node descriptors: no gateway-side plugin registration
required; gateway.nodes.pluginTools.enabled off-switch (default on);
description/count caps; deterministic node-prefixed collision names.
- Declarative node-hosted MCP: nodeHost.mcp.servers (McpServerConfig shape)
starts MCP clients on the node host, publishes tools as pluginId node-mcp,
executes via built-in mcp.tools.call.v1 with per-layer timeouts, failure
isolation, and orphan-safe shutdown. No re-pairing when servers change.
- Node-hosted skills: node.skills.update publishes ~/.openclaw/skills
content (64 skills/64KB/512KB caps both sides); gateway merges them into
the skills snapshot while connected and exec host=node is available, with
node:// locators, node-prefixed collisions, disabled command dispatch,
and gateway.nodes.skills.enabled + nodeHost.skills.enabled switches.
- Security: node-supplied pluginIds cannot satisfy pluginId-scoped tool
allowlists unless gateway-registered; reserved node-mcp id requires the
core MCP descriptor shape; protocol registry kept out of public
plugin-sdk dts.
- E2E: pond harness proves publication, MCP round-trip, skills locator, and
disconnect/reconnect for all three surfaces.
* style: format node-plugin-tools test
* fix(skills): keep status loader unfiltered when eligibility is passed
skills.status started passing eligibility for the node-skill merge, which
flipped loadWorkspaceSkillEntries into filtered mode and dropped disabled
skills from status reports (QA plugin-lifecycle-hot-reload timeout). Status
now merges node skills explicitly around an unfiltered load. Also: regen
docs_map for new node docs sections; add the intentional node-host MCP
onclose suppression to the lint-suppression allowlist.
* chore(lint): enforce no-floating-promises repo-wide
* feat(tooling): split DOM globals out of the Node-side typecheck program
tsconfig.core.json now compiles src/ + packages/ with lib ES2023; new
tsconfig.ui.json owns ui/ with DOM. Web-global names used by Node code
alias to canonical undici/stream-web types; highlight.js loads through a
validated narrow API because its d.ts force-includes lib.dom. Boundary,
sparse-guard, changed-lane routing, and profile lanes cover the new ui
graph.
* chore(deps): declare undici-types and teach knip about the highlight.js runtime require
* ci: retrigger dropped pull_request synchronize event
* refactor(types): derive web-global aliases from @types/node fetch globals, drop undici-types dep
* ci: retrigger after dependency guard cleared
* fix(scripts): add ui lane to changed-lanes declaration union
* chore(types): add declaration files for scripts/lib and scripts/e2e modules
* chore(types): add declaration files for top-level script modules (a-m)
* chore(types): add declaration files for top-level script modules (n-z)
* test: use a non-secret-shaped gateway token fixture
* test: type ci workflow guard helpers for the root test lane
* chore(tooling): typecheck root test/** with a dedicated tsgo lane
- test/tsconfig/tsconfig.test.root.json: root-test program (strict unused checks,
fixtures excluded; two Docker E2E clients that import built dist/** stay out,
same rationale as the scripts/e2e exclusion in tsconfig.scripts.json)
- tsgo:test:root wired into tsgo:test, check:test-types, scripts/check.mjs, and
the ci.yml test-types shard, mirroring the tsgo:scripts lane (#104348)
- changed-lane routing: test/**/*.ts (excluding fixtures) and the lane tsconfig
now trigger 'typecheck test root' in check:changed; previously test/ paths ran
lint only, so harness type errors surfaced first in CI (#104287 envDir case)
- burn down all 1071 latent type errors in the program: precise param/local
types across test/scripts, test/vitest, test/e2e, and transitive scripts/e2e
program members; 205 sibling .d.mts declaration files for imported .mjs
modules (committed separately); zero any, zero ts-expect-error
- resolve the pre-existing testing star-export ambiguity in
scripts/e2e/parallels/common.ts with an explicit re-export
Closes#104388
* chore(types): correct declaration fidelity per structured review
- re-derive 51 .d.mts files from implementation data flow instead of
initializers: fix a wrong never return (runTestProjectsDelegation returns
the child), add encoding-sensitive exec/spawn overloads (plain-gh), restore
the full release profile union, make parsed paths string | null, add missing
parseArgs fields via help/non-help unions, add a missing sibling declaration
(budget-number-args), drop 15 unused lint directives
- precise install-record/tuple typing removes the type-aware oxlint
regressions the first declarations caused in scripts/e2e implementations
- route .mts declaration edits under test/ to the testRoot lane and reference
the test-root project from tsconfig.projects.json so tsgo:all covers it
(closes both review findings against the lane wiring)
* chore(scripts): keep telegram runner dist typing structural for the boundary guard
* chore(types): declare runtime pack and gateway readiness exports added on main
* test: pin the importTargetPlan form of the plugin-contract plan import
The guard expectation still referenced the raw await import( form that
7ae5996bb3 (#103975) replaced with the importTargetPlan fallback helper;
the assertion fails on current main.
* origin/main:
fix(cron): abort superseded reconciliation hooks (#104368)
ci(release): expose Telegram runtime preflight stage (#104387)
fix(status): avoid false shell-wrapper audit warnings (#81778)
fix: retry live Gateway readiness proof (#104374)
fix(exec-approval): stop misattributing Allow Always unavailability to policy (#97740)
fix: preserve Mac bundle during live builds (#104376)
chore(docs): translate with GPT-5.6 xhigh
test: wire Control UI suite into CI, fix its broken tests, drop dead harness surface (#104361)
fix(ui): hide group submenu separator when New group is the only entry (#104370)
fix(discord): reset progress drafts across queued turns (#102341)
docs(cli): clarify that exec-policy show and approvals get exclude per-session /exec overrides (#94999)
fix(discord): single-source thread-binding default placement and guard artifact parity (#104342)
fix(agents): add tool-activity heartbeat to keep subagent alive during tool calls (#95536)
fix(maint): reuse recent same-PR hosted gates (#104355)
* test: wire Control UI suite into CI, fix its broken tests, drop dead harness surface
- fix 7 Control UI tests broken on main: module-mock factories are unreliable
under isolate:false shared workers; use RealtimeTalkSession prototype spies
and real jsdom storage instead (chat-realtime, chat-pull-requests)
- add gated checks-ui CI job (runUiTests changed-scope output): chromium
provisioning, lint:ui:no-raw-window-open, pnpm --dir ui test; ~30s on a
4vcpu runner, runs only when ui-affecting paths change
- add weekly node22-compat workflow exercising the supported lower Node bound
with the same command set as the dispatch-only ci.yml job
- delete the unreachable channels entry in EXTENSION_TEST_CONFIG_ROUTES and
the never-populated extensionRoutedChannelTestFiles override machinery
(born empty in 2ccb5cff22); resolved globs are provably identical
- guard run-vitest.mjs hardcoded path maps with existence tests so renames
cannot silently drop extended stall watchdogs
- remove unenforced coverage thresholds; test:coverage is informational, docs
updated (reference/test, help/testing, plugins/sdk-testing)
Closes#104321
* test: pin OPENCLAW_CI_RUN_UI_TESTS in the manifest env expectation
* test(ui): capture the realtime start spy instead of referencing the unbound prototype method
* test(ui): raise ui vitest timeouts for real-browser layout tests on small CI runners
* feat(tooling): add tsgo typecheck lane for scripts/**
* fix(scripts): burn down scripts type debt surfaced by the new lane
Typing-only except bugs the lane surfaced: gh-read timeout race,
Discord Headers spread dropping entries, undefined allowedHeadBranches
match, plugin-boundary matchAll crash. Deletes retired config keys from
fixtures/benches (prompt snapshots regenerated, config dump only) and
the orphaned non-runnable sync-moonshot-docs script. Adds full-surface
.d.mts declarations for existing .mjs boundaries.
* fix(watch): defer restart when dist/ is mid-rebuild (#99603)
Defer gateway:watch child restart when run-node's build/runtime-postbuild
readiness contract reports a hard dist/ failure. Only defers on conditions
where the child physically cannot start, preserving normal restart behavior
for soft staleness that run-node can rebuild on start.
Hard failures (defer restart):
- missing_dist_entry (mid-rebuild)
- missing_bundled_plugin_dist_entry
- missing_private_qa_dist
- missing_runtime_postbuild_output
Soft staleness (allow restart, run-node rebuilds):
- missing_build_stamp (w/ direct entry.js check for masked case)
- git_head_changed, dirty_watched_tree
- config_newer, build_stamp_missing_head, source_mtime_newer
isBuildReadyForRestart delegates to resolveBuildRequirement and
resolveRuntimePostBuildRequirement from run-node.mjs, using the
watcher process environment for consistency with the child runner.
200ms polling with 5-min timeout. Child-exit bounded recovery.
Clears stale restartRequested on timeout/poll errors.
Ref: #99603
Co-Authored-By: Claude <noreply@anthropic.com>
* fix(watch): deduplicate deferral loops, add single-flight guard, remove lint suppression
- Extract deferUntilBuildReady(onReady, onTimeout) shared helper,
replacing the near-identical pollBuildReady and pollAfterChildExit.
- Add single-flight guard (deferredRestartActive) so concurrent file
change events during deferral coalesce into one poll loop instead of
spawning a git-subprocess storm.
- Replace inline .catch arrow callback with method-style call, removing
the oxlint-disable-next-line suppression and its allowlist entry.
- Add tests: coalescing multiple changes, requestRestart recovery.
Ref: #99603
Co-Authored-By: Claude <noreply@anthropic.com>
* fix: keep gateway watch alive during concurrent builds
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Peter Steinberger <steipete@gmail.com>
* test: fix changed-run routing, dedupe gateway package lanes, repair force-rerun triggers
- route extensions/active-memory changed runs to their dedicated lane; add a
generic guard that every catch-all-excluded extension root resolves to a
dedicated config so future split lanes cannot silently skip changed tests
- exclude packages/gateway-client and packages/gateway-protocol from the unit
lane; explicit targets now route to the gateway-client lane (was double-run)
- generate forceRerunTriggers from the test/vitest directory and resolve all
entries absolute: Vitest matches triggers against absolute changed-file
paths, so the old hand-maintained relative list never matched at all
- replace deprecated envFile:false with envDir:false (Vitest 4.1.9 warning)
- give scripts/test-live.mjs stall detection teeth: kill the process group and
exit non-zero when OPENCLAW_VITEST_NO_OUTPUT_TIMEOUT_MS elapses quietly
- consolidate the live-docker shell target list into changed-lanes.mjs so the
lane regex and check targets cannot drift; cover subagent-announce edits
- delete orphans: vitest.extension-clickclack.config.ts,
vitest.full-core-unit.config.ts, scripts/test-docker-all.sh,
scripts/test-live-acp-spawn-defaults-docker.sh
Closes#104231
* test: keep envDir literal so sharedVitestConfig satisfies Vite UserConfig
Vite 8 types envDir as string | false; the bare object literal widened false to
boolean and broke check:test-types at the defineConfig call sites.
* feat: sidebar update card (web + macOS) with app-first mac update flow and Sparkle beta track
Squashed from claude/update-notification-display-c6cfb9 after semantic merge
with #104178 (channel-aware CLI installs). See PR #104171 body for details.
* chore(i18n): resync generated inventories after rebase
* chore(i18n): resync locale metadata after rebase
Adds a git-backed session diff panel to the Control UI, complementing the existing PR status chips.
New sessions.diff gateway method (operator.read): resolves the session checkout and returns structured per-file diffs (status, renames, +/- counts, capped unified patch) of branch + uncommitted + untracked work against the default-branch merge base. Hardened against git textconv execution and hardlinked out-of-tree content leaks; handles repos before their first commit via the empty-tree base.
Control UI renders a "Changes" panel in the chat detail sidebar with collapsible per-file diffs, hunk-gap markers, stat chips, and untracked/binary badges, gated on gateway method advertisement. Schemas additive (Swift models regenerated), 20 locales translated.
Closes#104182
* fix(install): harden stdin consumers to prevent pipe corruption in curl | bash
Redirect stdin from /dev/null for non-interactive subprocesses (npm install,
openclaw daemon restart, openclaw plugins update, openclaw dashboard) and
from /dev/tty for interactive ones (openclaw onboard in bootstrap). Also
protect the fallback paths in run_with_spinner and run_quiet_step.
This prevents subprocesses from consuming the script stream when the
installer is piped via curl | bash, which causes truncated function names
and hangs (reported in #73814).
Unlike the global pipe guard in #82918 (closed as too broad), this approach
has no detection heuristics and no re-execution. Each subprocess simply gets
the correct stdin for its purpose.
Fixes#73814
Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca>
* chore: retrigger proof evaluation
Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca>
* fix: redirect stdin in run_with_spinner raw-mode fallback
The success-path raw-mode fallback in run_with_spinner invoked the
command with inherited stdin. In a curl | bash scenario, this allowed
the child process to consume bytes from the script stream, causing
truncation. Add < /dev/null to match the other two fallback paths.
* retrigger proof check
* fix(test): update gum fallback assertion for stdin redirect
* fix: redirect gum-wrapped stdin and preserve TTY for interactive commands
Address ClawSweeper P1 and P2 findings:
- P1: Redirect stdin from /dev/null on the normal gum spin path so child
commands cannot consume the piped script stream (gum v0.17.0 passes
os.Stdin to wrapped commands).
- P2: Use is_non_interactive_shell to conditionally redirect stdin in
fallback paths. When running interactively (bash install.sh), commands
that need user input (e.g. Homebrew prompts) keep terminal stdin.
When piped (curl | bash), stdin is redirected from /dev/null.
- P2: Revert labeler.yml changes to keep the PR focused on installer
stdin safety. Size-label best-effort handling belongs in a separate PR.
Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca>
* chore: retrigger proof evaluation
Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca>
* retrigger proof check
* fix: base stdin isolation on stdin TTY check, not stdout
Replace is_non_interactive_shell with needs_stdin_isolation for stdin
redirection decisions. The new function checks stdin directly (! -t 0)
and NO_PROMPT, without checking stdout (-t 1). This ensures that
stdout redirection (e.g. install.sh > log.txt) does not suppress
interactive prompts when stdin is a terminal.
* test(install): add stdin isolation and NO_PROMPT coverage
Three focused tests for the needs_stdin_isolation function:
1. Verify needs_stdin_isolation returns true when stdin is piped
(the core curl|bash scenario).
2. Verify needs_stdin_isolation returns true when NO_PROMPT=1 is set
(explicit non-interactive override).
3. Verify run_quiet_step redirects subprocess stdin to /dev/null
when running in a piped context, preventing script consumption.
* test(install): strengthen stdin isolation test with sentinel data
Replace the weak TTY check (which passes even without the fix since
the test pipe is also non-TTY) with a sentinel-based test: pipe
SENTINEL_DATA on stdin and verify the child process reads nothing,
proving run_quiet_step actually redirects stdin to /dev/null.
* test(install): add counterproof and cat-based stdin isolation tests
Add two new tests to prove the stdin isolation fix is necessary and
works correctly:
- counterproof: demonstrates that pipe data DOES leak to the child
when stdin is not redirected through run_quiet_step, proving the
/dev/null redirect is the isolation barrier
- cat-based test: uses cat (reads all of stdin) instead of read -t 1
(timeout-based) for a deterministic assertion that run_quiet_step
produces empty stdin
* fix: gate gum spin stdin redirect on needs_stdin_isolation
The gum spin command unconditionally redirected stdin from /dev/null,
which also killed interactive prompts for direct installs since gum
v0.17 passes os.Stdin to the wrapped command. Now only redirect
stdin when needs_stdin_isolation returns true (piped install context).
Adds focused tests verifying both paths: piped installs get /dev/null
redirect, direct interactive installs preserve terminal stdin.
* test: assert gum stdin is not /dev/null for direct interactive installs
The direct gum runtime test now reads the child command's stdin-source
log file and asserts stdin was NOT /dev/null, using device:inode
comparison (stat -f on macOS, stat -c on Linux) for reliable detection
across both platforms.
* retrigger proof check
* retrigger proof check with real installer evidence
Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca>
* fix(install): preserve interactive post-install stdin
* fix(install): route piped prompts through controlling tty
* fix(install): keep quiet piped steps noninteractive
* fix(install): avoid hidden prompts with redirected output
* fix(install): preserve visible prompt output state
---------
Signed-off-by: Sebastien Tardif <sebtardif@ncf.ca>
Co-authored-by: Vincent Koc <25068+vincentkoc@users.noreply.github.com>
