A Gateway timeout or closed connection now fails the command with an
actionable stderr hint instead of silently re-running the whole turn
embedded under a fresh gateway-fallback-* session. The silent fallback
could double-execute side effects (the Gateway may still finish an
accepted turn), returned context-free answers to --session-key callers,
and ran with the CLI host's local config. --local remains the only
embedded execution path.
Also deletes the resultMetaOverrides plumbing (the fallback was its only
writer) and the fallback marker fields added in #111645.
Channel plugins declare package env triggers in manifests, and ambient
environment variables count as "channel configured". A dev gateway
(gateway run --dev) inherits the operator's shell, so real channel
credentials silently configured channels and could connect development
instances to live services (observed with reef and a Telegram-range
connection during isolated stress testing).
Dev mode now drops presence signals whose only sources are env or
manifest-env across activation planning, auto-enable, autostart, health
recovery, readiness, reloads, and startup warnings. Explicit
channels.<id> config still works, --dev-ambient-channels restores the
old behavior, and startup logs the suppressed channel ids once. Non-dev
gateways are unchanged.
* fix(plugins): load newest managed generation when a prior install lingers
An upgrade writes a plugin's new version into a distinct managed project
directory (an `__openclaw-generation__` dir) without removing the previous one.
The persisted install record keeps pointing at the older, still-present
directory, so `mergeRecoveredManagedNpmRecord` falls through to returning the
persisted record and the runtime imports the stale version even though the newer
generation is installed. `isUnavailableManagedNpmInstallRecord` only self-heals
when the persisted path is gone, so an upgrading install (both directories
present) is not covered — `plugins inspect <id> --runtime` reports the old
version and `install`/`update` appear to silently no-op.
- Recovery now selects the highest-version record per plugin id across the
legacy flat dir and any generation dirs, instead of whichever project root
sorts last (generation-dir suffixes are hashes, so on-disk order is unrelated
to version).
- When the recovered managed install is a strictly newer version at a different
path, repoint to it, reusing the metadata merge the unavailable-path branch
uses. The repoint is gated to persisted records that live inside the managed
npm root, so intentional custom/outside npm installs are left untouched.
This is a recovery-layer self-heal, so it also repairs installs that already
diverged from an earlier upgrade.
Refs #107228
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(plugins): honor active managed generations
Recover ambiguous managed installs by package recency only when no usable active ledger path exists, and let doctor retire non-authoritative generations safely.
Co-authored-by: Peter Lindsey <peter@lindsey.jp>
* test(plugins): use shared temp cleanup
* fix(plugins): compare active paths per platform
* fix(plugins): use managed install timestamps
* test(plugins): timestamp managed project roots
* test(plugins): exclude retired legacy recovery
* refactor(plugins): keep recovery candidate private
---------
Co-authored-by: Peter Lindsey <peter@lindsey.jp>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Peter Steinberger <steipete@gmail.com>
* fix: bound misc unbounded fs.readFile calls; remove unused fs import
* fix: decode buffer to string before passing to string consumers
readRegularFile and readRegularFileSync return { buffer, stat },
not a string. All 4 new call sites passed the raw object to functions
expecting a string (JSON.parse, RegExp.test, template literals, etc.),
causing TS2345 type errors and runtime failures.
Fix each call by extracting .buffer and calling .toString('utf8')
before passing the result to string consumers.
* style: fix oxfmt formatting in config-set-input.ts
* fix: bound config and trajectory metadata reads
Co-authored-by: 陈宪彪0668000387 <chen.xianbiao@xydigit.com>
* refactor: isolate bounded read ownership
Co-authored-by: 陈宪彪0668000387 <chen.xianbiao@xydigit.com>
---------
Co-authored-by: Peter Steinberger <steipete@gmail.com>
* fix(gateway): emit config.changed from the reload committer for every accepted write
Agent config_set, CLI, and doctor writes reach the gateway through the file
watcher's reload path, which never broadcast config.changed - only direct RPC
writes did. Move the broadcast to a single onConfigCandidateCommitted hook in
the reload committer so all writers notify connected UIs, including
runtime-skipped commits (writer-intent echo suppression, reload mode off),
and delete the duplicate RPC-side broadcasts.
* feat(ui): sync chat follow-up and commentary prefs across devices
Add chatPersistCommentary and chatFollowUpMode to the ui.prefs sync surface
(schema, types, docs) with clearable null removal for the follow-up override.
Rewrite server-prefs around a descriptor table so a synced pref is one entry
instead of five hand-written code sites. Caption synced rows/sections
(chat prefs, theme, language) and refresh the stale Settings subtitle.
* refactor(ui): dedupe settings constants, drop dead composer branch, refresh copy and docs
Derive the Security tool-profile options from PROFILE_OPTIONS and General's
thinking levels from BASE_THINKING_LEVELS instead of parallel literals.
Remove the unused composer mode:"model" branch, the orphaned
.config-view-toggle CSS family, and stale quick-settings wording. Add static
settings-search entries for the Sessions and Managed Worktrees pages plus a
provenance-row test, and update docs/cli/openclaw.md to describe the current
system-agent config-write policy (denylist roots, route-backing plugin guard,
per-agent routing escalations).
* chore: refresh plugin SDK API baseline after rebase onto current main
* fix(secrets): use resolveStateDir for .env path to cover legacy .clawdbot dir
* fix(secrets): replace override-based legacy .env test with true automatic fallback, add apply-path coverage
* fix(config-cli): emit JSON error on config get --json for missing paths
When is used with a nonexistent
path, the command emits no JSON on stdout, only a text error on stderr.
This breaks automated consumers that expect machine-readable errors.
Fix: output JSON to stdout before exit(1) when
opts.json is set. This matches the existing JSON error output pattern
already used by runConfigValidate and handleConfigMutationError.
Also re-throw synthetic exit errors in the catch block so the
--json branch's writeRuntimeJson+exit is not caught and re-emitted
as a text error.
* fix(config-cli): use typed ExitError signal instead of __exit__ string sentinel
* fix(secrets): reuse resolved stateDir for .env path in audit
* fix(secrets): pass resolved stateDir into apply .env scrubbing, add root-switch regression
* test(config-cli): migrate remaining __exit__ sentinel to ExitError
* fix(secrets): cover config and state dotenv roots
* style(secrets): format dotenv path helper
* test(secrets): track temporary directories
* test(config): expect typed exit for missing patch files
---------
Co-authored-by: Peter Steinberger <steipete@gmail.com>
Co-authored-by: Peter Steinberger <peter@steipete.me>
* perf(sqlite): gate schema version before integrity scans and quarantine terminal open failures
* feat(gateway): refuse incompatible database schemas at boot and verify integrity in the background
* feat(update): refuse installing builds that cannot open the current databases
* docs: add database schema reference with version history and downgrade guidance
* test(update): tolerate optional runner options in git-mutation mocks
* refactor(update): split npm package-target metadata out of update-check
* chore: model verifier internals for knip and regenerate docs map
* chore(i18n): resync stale iOS localization catalog
* chore: allowlist read-only preflight and verifier raw SQLite access
* feat(onboarding): persist app recommendations once
* feat(onboarding): add self-naming birth sequence
* chore(wizard): hoist bootstrap-defer imports to the top of the module
* docs(bootstrap): persist agreed identity into IDENTITY.md and SOUL.md as well
* docs: regenerate docs map after onboarding plan merge
* fix(wizard): reuse a pending stored offer instead of rescanning apps
* fix(onboard): deduplicate bootstrap recommendations
* feat(onboard): add browser-first guided handoff
* chore: drop release-only CHANGELOG edit from browser-handoff branch
* fix(onboard): make browser handoff reach the gateway on real installs
The presence probe connected as a raw shared-auth client with a possibly
SecretRef-managed token, which the gateway rejects as an unpaired Control UI
client ("device identity required"), so the handoff never fired and always
fell through to the terminal hatch. Read presence as a CLI-mode loopback client
(the trusted path every openclaw command uses) so operator.read is granted, and
share that one auth path between the reachability gate and the wait loop. Also
short-circuit when gateway.controlUi.enabled is false so onboarding does not open
a dead dashboard URL and block for the full timeout.
* fix(onboard): share resolved target between handoff gate and wait
The reachability preflight now resolves the same target (and shared secret) the
wait loop uses, and presents the configured secret to the CLI-mode presence
read. Known remaining limitation: on fresh quickstart gateways whose auth token
is written to config but resolved differently at runtime, the presence read is
rejected with a token mismatch and the handoff falls back to the terminal hatch.
Tracked for follow-up before enabling the browser handoff by default.
* docs(onboarding): mark phase 3 browser handoff as held with diagnosis
* docs(onboarding): phase 3 proven end-to-end; token-mismatch hold was a test artifact
* fix(onboard): satisfy docs-map and test-type gates for the handoff branch
* refactor(onboard): split guided manual stage
* feat(onboarding): recommend plugins and skills from installed apps
Scan installed macOS apps during classic onboarding (TCC-free), gather
candidates from official catalogs + ClawHub search, let the configured
model pick genuine matches, and offer an opt-in multiselect install step.
Adds a device.apps node-host command (default-off sharing, Android-parity
envelope) so remote gateways can request a paired Mac's inventory, and a
wizard.appRecommendations kill switch. Custom setup-inference completions
no longer inherit the 32-token verification-probe output cap.
* feat(onboarding): recommend apps in guided flow
* fix(onboarding): harden app recommendations against ClawHub self-promotion
Third-party ClawHub skills are never pre-selected regardless of model tier
(publisher-controlled listing text reaches the matcher prompt and could
promote itself); their labels now say they install third-party code.
Installed-app scans follow symlinked .app bundles. Matcher output stays
bounded by the resolved model's own maxTokens budget (documented invariant).
* fix(onboarding): key official catalog candidates by resolved plugin id
Real catalog entries are package manifests without a top-level id; keying the
candidate map and channel/provider classification by entry.id collapsed the
whole official catalog into one undefined-keyed entry, so no official plugin
or channel was ever recommended. Regression test runs against the bundled
catalogs.
* fix(onboarding): satisfy lint, types, deadcode, and migration gates
Split the guided-onboarding test into a self-contained custodian suite to stay
under max-lines. Narrow app-recommendation exports (drop dead node-payload
normalizer, unexport internal types/helpers, route candidate tests through the
public API), replace map-spread with a helper, unexport device.apps result
types, add installedAppsSharing to node-host migration expectations, cast the
wizard multiselect mock, and regenerate the docs map.
* test(onboarding): register new live test in the shard classifier
