ensureGatewayProfile used to activate the target gateway and set
$activeGatewayProfile while the connection descriptor fetch was still
in flight, so during that window $gateway already targeted the new
backend while $connection still described the previous one, and any
request or plugin mode-listener firing then announced the wrong mode to
the new backend. A failed descriptor fetch made the mismatch permanent.
prepareGatewayForProfile (new gateway-store seam) opens the socket and
returns a synchronous activation thunk without publishing anything;
ensureGatewayForProfile now delegates to it. The switch resolves the
descriptor and opens the socket first, then flips the active gateway,
the profile atom, and $connection in one synchronous frame. A
descriptor failure aborts the switch as a unit: nothing is published and
every atom still consistently describes the previous profile.
The deferred-descriptor test holds the fetch open and asserts the public
atoms never disagree, then releases it and asserts all three flipped
together; the failure test asserts no partial publication.
Adds a reviewable in-app install path for Hermes plugins:
hermes://plugin/install?repo=owner/repo (and Settings -> Plugins ->
Install from Git) opens a confirmation modal showing the repo identity
and source links, shallow-clones to probe for agent and/or desktop
plugin artifacts, lets the user pick components, then installs — agent
side through the gateway's new plugins.manage `install` action (wrapping
the existing dashboard_install_plugin), desktop side through a new
Electron git-install module with subdir-escape guards, a 60s clone
timeout, non-interactive git env, and insecure-scheme warnings. Never
auto-installs; hybrid repos get one dialog. Legacy plugin-agent /
plugin-desktop deeplinks route into the same modal.
Salvaged from PR #82735 by @serefyarar (net diff applied onto current
main as a single authored commit; the branch carried merge commits).
The preview screenshot PNG from the original branch was intentionally
not carried over — images live in PR bodies, not the repo.
The revert to the committed theme lagged behind Escape. The palette
body stays mounted through the whole exit animation, and the preview
was cleared at unmount. So the repaint waited for the fade.
Subscribe to the palette open store in the body and clear the preview
the moment the store flips to closed. The unmount clear stays as the
backstop for a body that dies without a close.
The highlight preview did not fire. cmdk calls the root onValueChange
only in controlled mode, when the value prop is set. The palette is
uncontrolled, so the preview callback never ran.
Add a HighlightWatcher child that subscribes to the cmdk store with
useCommandState. The store reports the highlight in both modes. The
watcher replaces the dead root prop.
The new test renders a real uncontrolled cmdk root. It proves that
the watcher fires and that the root prop stays silent. If cmdk later
fires the prop in uncontrolled mode, the second assertion fails, and
the watcher becomes removable.
The theme rows in the Cmd-K picker applied a theme only on select.
Now the highlighted row paints its theme immediately.
cmdk reports the highlighted row through onValueChange on the root.
A new optional onHighlight callback on PaletteItem receives it. The
theme rows preview through a new previewTheme function on the theme
context. The preview is not persisted. A highlight on a row without
onHighlight, a page change, a palette close, or a commit clears the
preview. Then the committed appearance returns.
CI slice 11 caught a regression from the salvaged get_provider() fallback:
the "custom" placeholder profile (aliases ollama/local/vllm, empty
base_url) now resolved as a bare ProviderDef before
resolve_provider_full() reached its custom_providers step, collapsing
keyed IDs like custom:local-127.0.0.1:11434 to an endpoint-less "custom"
(test_keyed_custom_provider_bare_custom_fallback_uses_stable_key).
Gate the fallback on the profile carrying a concrete base_url —
placeholder profiles completed by config.yaml keep flowing to the
custom-provider resolution path.
Follow-up to the salvaged CommandCode signature fix: accepting base_url
but ignoring it left custom endpoints (user-configured model.base_url /
COMMANDCODE_BASE_URL proxies) fetching the public catalog instead of the
configured one. Reviewer dansigma flagged this on PR #88851.
Class-wide fix, not a CommandCode patch:
- providers/base.py: a caller base_url that DIFFERS from the profile's
default now wins over models_url. Equality with the default means "not
customised" (callers pass base_url unconditionally, defaulting to the
profile's own URL) and keeps models_url as the endpoint, preserving the
OpenRouter-style split-catalog behavior.
- commandcode: _fetch_commandcode_models() takes the endpoint override;
both profile overrides forward base_url.
- Tests: base-class precedence (custom beats models_url, default does
not), CommandCode redirect via live local HTTP server incl. claude-*
filter, and default-echo hitting the canonical endpoint. All verified
to fail against the pre-fix implementation (sabotage run).
Plugin-only providers (commandcode, tencent-tokenhub, ...) are absent from
models.dev and HERMES_OVERLAYS, so resolve_provider_full returned None and
/model switches failed with "Unknown provider ..." even though the picker
lists them (CANONICAL_PROVIDERS auto-extends from the same registry).
Fall back to providers.get_provider_profile() before giving up, mapping the
profile api_mode to the ProviderDef transport.
The model picker's generic live-fetch path (hermes_cli/models.py
provider_model_ids) calls profile.fetch_models(api_key=..., base_url=...).
Both CommandCode overrides only accepted api_key/timeout, so every picker
open raised TypeError, which was silently swallowed, leaving the provider
with zero models.
Match the base ProviderProfile.fetch_models signature (base_url kwarg) and
add a regression test asserting both profiles accept it.
The global SESSIONS sidebar only aggregated local profiles and v1 per-profile
remote overrides. Sessions living on v2 registry connections (remote/cloud/ssh
gateways) never appeared — the remote API returned the rows, but the renderer's
Sessions component received an empty array (#88880).
- electron/profile-session-routing.ts: fetchRegistrySessionRows reads each
CONNECTED registry gateway's session list (ssh backends natively, shared
remote/cloud hosts via one cross-profile aggregate with a legacy flat-list
fallback), tagging rows with connection_id + owning profile.
spliceRegistrySessionRows dedupes them into the unified list and extends
per-profile totals. Reads never pass include_hidden, so Bot Mode's hidden
canonical chats stay OUT of the global list, same as local sessions.
- electron/main.ts: mergeRemoteProfileSessions splices registry rows; the
/api/profiles/sessions[+/sidebar] intercepts also fire when registry
gateways are pooled (previously only v1 remote overrides). Only
already-pooled backends are read — a sidebar refresh never dials or spawns
a backend (the roster-respawn trap), and a dead gateway contributes nothing.
- types/hermes.ts + use-session-actions: SessionInfo carries connection_id;
resuming a registry-owned row activates its connection-scoped gateway
(ensureGatewayAgent) instead of a same-named local profile.
- store/gateway.ts: ensureActiveGatewayOpen rides out an in-flight secondary
activation (bounded 8s) instead of failing instantly — the Sessions "+"
during remote wake no longer errors "Hermes gateway is not connected".
Tests: 5 new registry-source/splice unit tests (tagging, shared-host
aggregate + legacy fallback, dead-gateway isolation, hidden-flag contract,
dedupe/totals) and a sabotage-verified activation-wait regression test.
Two hardening follow-ups on the salvaged #84192 work:
- dispatchNativeNotification now reports whether the notification actually
reached the OS bridge, and dispatchPluginNativeNotification registers its
onActivate/onAction closures only on true. Previously a throttled,
disabled, or baseline-suppressed notification registered handlers that no
click could ever clear, leaking them for the window's lifetime.
- The renderer's onNotificationActivate handler re-resolves the activate
payload through resolveHermesOpenPath instead of trusting the pre-IPC
validation, keeping path validation in one funnel for any future
hermesDesktop.notify caller.
Adds a regression test covering the throttled and suppressed cases.
Extends ctx.os.notify (the curated plugin OS door from #78685) with icon,
action buttons, and a serializable `activate` target. Body/action clicks
focus the window and navigate to the plugin's screen; activation paths
share one resolver (hermes-open-target.ts) with hermes:// OS deep links,
so `hermes://index-network/intent/1`, `/index-network/intent/1`, and
{ path, params } all land on the same hash-router route. Approval
notifications keep their existing session-scoped channel.
Salvaged from PR #84192 by @serefyarar (net diff of the PR branch applied
onto current main; branch carried merge commits so a single authored
commit preserves attribution).
PR #67934 marked auto-discovered catalogs by writing two sentinel keys
INSIDE the user-facing ``models`` mapping of custom provider entries:
``__discovered_model_catalog__`` (written by
_save_discovered_models_to_config) and ``__explicit_model_allowlist__``
(injected by _normalize_custom_provider_entry). Every consumer of that
mapping — pickers, selectors, gateway/agent readers, and the user's own
config.yaml — had to know to filter those keys, and any site that
didn't listed them as phantom model IDs (``__discovered_model_catalog__``
showing up as a selectable "model"). The v11→v12 config migration and
the ACP session-state test caught exactly that leak on main.
Replace the in-mapping sentinels with a single entry-level flag:
- ``models_discovered: true`` now sits next to ``models``/``base_url``
on the provider entry; the models mapping stays a clean
``{model_id: metadata}`` dict with no reserved keys.
- _save_discovered_models_to_config writes the new shape and refreshes
catalogs it previously discovered (entry-level flag or legacy
sentinel) instead of treating them as user-curated metadata.
- _normalize_custom_provider_entry no longer injects
``__explicit_model_allowlist__``; a dict-shaped models mapping counts
as an explicit allowlist exactly when the entry is NOT marked
models_discovered.
- _models_config_is_allowlist takes the discovered flag as a parameter
(new helper _entry_models_discovered resolves it, including the
legacy in-mapping sentinel); all call sites updated
(model_switch.py, model_setup_flows.py, acp_adapter/server.py).
- Backward compat, no config version bump: configs written by a
pre-fix Hermes (sentinels inside models) still read correctly —
``__discovered_model_catalog__: true`` is treated as
models_discovered, both sentinel keys are stripped from model
listings, and the next discovery save migrates the entry to the
clean shape. Covered by a new regression test.
Also restore ``except Exception:`` on the pre-existing guards this PR
had narrowed to specific exception tuples (the resolve_runtime_provider
fallback in switch_model, the picker discovery/cache guards in
list_authenticated_providers, _get_model_config_dict, and
_credential_fingerprint). Those guards were intentionally broad on
main — a failed resolution or probe must degrade to the fallback path,
never crash the model switch. Guards the PR introduced for its own new
probe code keep their authored tuples.
The ACP new_session payload also goes back to
probe_current_custom_provider=False, matching the contract main's
test_new_session_returns_authenticated_cross_provider_model_state pins
(session opens must not block on live-probing the current custom
endpoint).
The inline branch of MATH_SPAN_SPLIT_RE excluded `$` from the body outright,
so a `\$` inside inline math — a literal dollar sign, valid TeX — broke the
span match and the shield silently didn't apply. `$\sqrt[3]{8} + \$5$` still
lost its index.
Step over escape pairs instead, matching the escaped-delimiter rule
findClosingSingleDollar already applies via isEscapedAt. The two body
alternatives are disjoint on their first character, so the added quantifier
can't backtrack ambiguously.
Reported by Copilot in review.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SPGjEQ2yrS4nWiooUEdYti
`$\sqrt[3]{8}$` renders as a plain square root — the index is gone. It is
not a KaTeX layout problem: the index never reaches KaTeX. CITATION_MARKER_RE
strips `[3]` as a citation marker, because its lookbehind accepts any letter
and the `t` of `\sqrt` qualifies. That runs inside normalizeVisibleProse,
which splits out inline code spans but not math, so TeX is fed to rewrites
written for prose.
Numeric-only, which is why `\sqrt[n]{8}` survives and made this look like a
layout edge case rather than a preprocessing one.
Shield math the same way inline code is already shielded: split each prose
part on math spans and rewrite only the segments between them. That also
takes math out of the reach of the other rewrites in that pass
(autoLinkRawUrls, LOCAL_PREVIEW_URL_RE, the ``` stripper, linkifySessionRefs),
any of which can corrupt TeX the same way with different input.
The split is capturing, and math segments are identified by index parity
rather than a leading `$`, so a prose run that merely opens with a stray
dollar cannot be mistaken for math. Escaped `\$` delimiters stay prose, which
is what keeps `$5 and $10` escaping intact.
Verified in the real renderer through the desktop mock-backend E2E harness:
`.katex .root` (the span KaTeX emits for a radical index) goes from 1 to 4 on
the same four-radical reply, and the MathML annotations show KaTeX receiving
`\sqrt[3]{8}` intact rather than `\sqrt{8}`.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SPGjEQ2yrS4nWiooUEdYti
Multi-line display math whose $$ delimiters hug the body
(e.g. $$\begin{aligned}...\end{aligned}$$) renders as raw error text.
remark-math's flow-math construct is fence-shaped: text after the
opening $$ on the same line is read as an info string and discarded,
and the closing $$ is only recognized alone on its own line. So the
block never closes and KaTeX paints the remains via its error fallback.
splitHuggingDisplayMath moves those delimiters onto their own lines. It
runs AFTER normalizeMathDelimiters because that rewrite is itself a
source of the hugging form: a multi-line \[...\] comes out of it as
$$\begin{aligned}...\end{aligned}$$, so the same bug reached users who
never typed a $$ at all.
Single-line $$...$$ is left alone (it routes through the inline
math-text construct and already renders), container prefixes are
replayed onto the delimiter lines, and both patterns anchor $$ to the
start of the line, which keeps them from firing inside an inline code
span.
Verified end to end: the repro emits katex-error through
remark-math + rehype-katex before this change and not after.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
tools/skills_sync.py bound HERMES_HOME / SKILLS_DIR / MANIFEST_FILE at
import time — the third module in the same lineage as skills_tool
(f8723c478) and skill_manager_tool (c6a3d412d). In a long-lived
dashboard/TUI process, console skills commands (reset, diff,
list-modified, opt-in/out, repair-official) dispatched in-process under
_profile_scope's set_hermes_home_override(), but skills_sync's frozen
constants kept resolving against whichever profile was live at import.
Sharpest edge: reset_bundled_skill()'s #48200 rmtree strict-child guard
was computed against the WRONG skills root.
Fix: same call-time accessor pattern as the two prior fixes —
_hermes_home()/_skills_dir()/_manifest_file() honor an explicitly
patched module global (tests, retargeting) and otherwise re-resolve
from the live profile-scoped get_hermes_home() on every call. All 37
call sites migrated; module constants kept for compat.
Also documents in _profile_scope() that skills_sync needs no module
retargeting since the contextvar override now reaches it.
Regression tests (sabotage-verified: all 3 fail on the old binding):
- accessors follow set_hermes_home_override at call time
- explicit module patch still wins over the override
- rmtree guard anchors on the overridden profile's skills root
Fixes#65828
Bot Mode wake-ups died on a routing split-brain: session.resume /
session.activate / session.usage were dispatched on whatever socket was
active at request time, while the bot's own backend sat healthy and idle
(zero traffic until the idle reaper killed it). Two divergence sources,
both fixed:
1. Registry-owned route truth. applyActive() now publishes the active
route's bare profile ($activeGatewayRoute + onActiveRouteChanged), and
use-gateway-boot mirrors it into $activeGatewayProfile. Previously,
eviction fallbacks (idle reap, connection removal, profile delete)
moved the SOCKET back to the primary while the profile atom kept
naming the evicted bot — ensureGatewayProfile's "already active" fast
path then trusted the stale atom and skipped the re-swap forever.
2. Request-time routing for session-scoped RPCs. resumeSession's RPCs go
through requestForSessionProfile (store/session-request-router.ts):
when the active route serves the session's owning profile the ambient
dispatcher is kept (reauth-aware reconnect); when it diverges — a
concurrent switch won the mutex, a failed dial left the old socket
active, an eviction re-pointed the route — the RPC is pinned to the
owning profile's own socket via requestGatewayForProfile.
Diagnosed from zero trust's debug bundle (loki/hulk/teknium-kun backends
READY then idle-reaped, renderer stuck on "Waking up… → retries gave
up") and DanBennettUK's #89206 trace (profile socket accepts, closes
with messages=0, no resume RPC observed).
Both layers sabotage-proven: reverting the route publish fails the
lockstep/eviction tests; reverting the request-time routing fails the
wrong-socket dispatch test.
Assistant messages that link a file the agent wrote —
[report](/home/user/report.md), file://…, ~/…, C:\… — rendered as dead
anchors: file:// is blocked in the renderer, Streamdown's URL hardening
turns file:/~/ hrefs into "[blocked]" spans, and on a remote gateway the
path isn't on the viewer's disk at all. Issue #82140 proposed exposing
the Desktop connection mode to skills/MCP/plugins so EXTENSIONS could
emit different output per viewer; this fixes the symptom at the right
layer instead — the viewer surface resolves paths at VIEW time, so
extension output stays surface-agnostic and the same transcript works
from every machine that opens it.
- markdown-preprocess: routeFileLinksToPreview() rewrites filesystem-path
links in prose to the renderer's existing hash-href doors —
#preview/… (PreviewAttachment) for documents, #media:… for
audio/video/image extensions. These pass URL hardening by design and
resolve through normalizeOrLocalPreviewTarget / resolveMedia*Src:
local connections read the file directly, remote connections fetch
over the authenticated /api/fs bridge. Image syntax, fences, inline
code, anchors, relative and http(s) links untouched.
- markdown-text: MarkdownLink routes any filesystem href that still
reaches it (bypassing preprocess) to PreviewAttachment/MediaAttachment
instead of a bare dead <a>.
- media.ts: export isFileMediaPath.
Live E2E (built app, CDP-driven, fixture session with links to a real
gateway-side file):
- BEFORE: [report.md] = dead <a href="/home/…"> (click: nothing),
[notes](file://…) = "notes [blocked]" span, 0 preview affordances.
- AFTER: both render as attachment rows; Open preview shows the file's
content in the preview pane; zero blocked spans; screenshots verified.
Closes#82140. Supersedes PR #82187 (connection-mode API): with view-time
resolution the extension layer no longer needs to know where the viewer
sits.
A `.md` delivered via MEDIA has no entry in MEDIA_BY_EXT, so mediaKind()
classified it as a generic 'file' and MarkdownLink rendered a download-style
anchor. Markdown is renderable content: route markdown document paths to
PreviewAttachment (source='tool-result'), which opens them in the right-rail
preview pane — where .md already renders with a rendered/source toggle and,
since #89381, full KaTeX math, tables, images, and links.
Resolves#84951 (the MEDIA delivery half; the rail-side rendering half landed
in #89381).
The Bots pane docks into the sessions zone but did not declare
collapsible. Below the sidebar-collapse breakpoint, the sessions pane
left the grid and the zone kept a stranded BOTS tab on screen.
The Bots pane now declares collapsible, so it leaves the grid with
its zone. The narrow edge overlay now mirrors the zone's tab strip
when the revealed pane has collapsed zone-mates. Without the strip,
only the first pane of the zone was reachable while collapsed. A
lone pane keeps the stripless overlay form.
Each closeable horizontal tab now shows a close button when the
pointer is on the tab. A small gradient fades the button into the
tab surface, so long labels fade under it instead of a hard clip.
The gradient reads the tab's effective surface color. This color
tracks the hover and selection washes, so the fade is correct on
every theme.
Middle-click and Cmd-click still close the tab. Vertical rail tabs
keep those gestures and do not get the button.
_write_manifest still used a hand-rolled mkstemp + atomic_replace,
so every sync reset .bundled_manifest to mkstemp's 0600, dropping a
group-readable or shared mode the operator had set. Replace the block
with utils.atomic_write_text(preserve_mode=True) — the same shared
writer and mode-preservation contract PR #86255 applied to the skill
manager's document writes.
This is the remaining half of PR #14410 by @sgaofen, who reported the
manifest mode reset first; the skill-manager half of that PR was
superseded by the atomic_write_text refactor and #86255.
Export MarkdownPreview and cover the regression this PR fixes: KaTeX output present with no raw $$ delimiters left, GFM table structure, image alt/src, and external links opening with noopener noreferrer. Aligns the import order (perfectionist/sort-imports) and documents the mathPlugin module-scope setup to match the chat renderer.
The preview renderer for .md files was missing the math plugin, table/image/link components, and the markdown preprocessing pipeline that the chat transcript renderer has. Add KaTeX math rendering (inline $...$ and block $$...$$), table, image, and link support so file previews match the chat rendering.
DOM-level render test mounting the exact file-preview pipeline
(normalizeFilePreviewMath -> Streamdown + memoized math plugin) and asserting
`.katex` output for inline $.x.$, display $$..$$, and \(..\) delimiter math,
plus that a code-fence $.HOME.) stays code.
The right-rail file preview rendered markdown through Streamdown with no
plugins and no math preprocessing, so $...$ and \(...\) stayed as raw
source text. Wire the memoized KaTeX plugin (same one the chat transcript
uses) into the preview and preprocess prose with a math-only normalizer that
skips chat-only transforms (reasoning-block stripping, session-ref linking,
preview-target stripping, URL autolinking, citation stripping) so a file's
prose, code fences, and inline code spans are never mangled.
- preview-file.tsx: import + module-scope createMemoizedMathPlugin, pass
plugins={{ math }} to Streamdown, preprocess text with
normalizeFilePreviewMath before render
- markdown-preprocess.ts: add exported normalizeFilePreviewMath
- markdown-preprocess.file-preview.test.ts: 8 tests (currency escaping,
delimiter normalization, fence/inline-code preservation, verbatim
citations/URLs/reasoning blocks)
Follow-up to #89382: the operator runbook, bundled-skill docs page, and the
bundled SKILL.md now cover the organizer-scoped lookup flag and note that
/meet/ short URLs require it while webhook jobs derive the organizer
automatically.
_fetch_owner_handle and the browse catalog walk both inlined the
same owner-handle extraction logic that _owner_from_payload was
extracted to centralize. Replace both with calls to the helper.
ClawHub treated the last path segment as a slug, so a GitHub-style
id like owner/repo/skills/skillopt fetched a different author's
skillopt. Pair metadata and files from the same source so inspect
cannot show one registry's header and another's SKILL.md.