Files
Vyctor H. Brzezowski aba94bbe0b fix: install or review the publisher you picked when ClawHub skills share a slug (#121697)
* fix(skills): keep ClawHub publisher identity from search through install

ClawHub search returns one entry per publisher, so several results can share a
slug. Every client collapsed the selection to that bare slug before calling
skills.detail and skills.install, and ClawHub answered 409 AMBIGUOUS_SKILL_SLUG
with no in-product way forward.

searchClawHubSkills now records the publisher-qualified reference once, on the
result that carries it, and the Gateway protocol documents it. skills.detail
parses the same reference grammar skills.install already accepted, so review and
install cannot resolve to different publishers. Control UI carries that one
reference through row actions, detail, busy state, and acknowledgement retries,
and shows it so otherwise identical rows are distinguishable.

Fixes #117633

* fix(apps): send the ClawHub publisher reference from native skill browsers

macOS, iOS, and Android read the qualified reference from search results and use
it for skills.detail, install, busy state, installed matching, and list identity,
so two publishers sharing a slug stay distinct instead of collapsing into one
ambiguous request.

* fix(skills): refuse external-source skill detail instead of reading a same-slug skill

ClawHub has no source-qualified read endpoint, so a skills-sh reference parsed
down to its bare slug would have returned a registry skill's card while install
resolved the external artifact. Review and install could name different skills.

skills.detail now fails closed on any reference that carries a source, and the
macOS and AgentPro rows show the publisher reference next to the summary instead
of only when a summary is missing, so same-slug rows stay distinguishable.

* chore(apps): refresh native i18n source baseline for the skill row references

* refactor(skills): drop the unread search-result ownerHandle field

installRef is the one reference clients send back, and no client reads the
publisher handle separately, so the protocol and Control UI carry one field
instead of two.

* fix(skills): name the next step when external skill detail is refused

Clients that gate install behind a successful review would otherwise see only a
refusal, so the error names the direct install path and the CLI equivalent.

* fix(macos): use a doc comment on the ClawHub row subtitle

swift-format's docComments rule requires doc comments on declarations; the
subtitle property carried a regular comment and failed macos-swift.

* fix(skills): carry ClawHub trust state to clients that can install

Forwarding installRef let clients install the exact publisher the operator
picked, including external skills-sh sources. It did not forward the trust
state that says ClawHub never scanned that source, so iOS AgentPro — the one
surface that installs in a single tap with no review step — could install an
unscanned artifact with nothing on screen saying so. The CLI already labels
these (docs/clawhub/cli.md, docs/cli/skills.md); native clients could not,
because trustState was never on the wire.

trustState becomes an optional field on SkillsSearchResultSchema. It is purely
additive: older clients ignore an unknown key and the field is absent for
registry results, so downgraded readers are unaffected and no protocol version
moves.

Every client that renders a search row now shows "Not scanned by ClawHub",
matching the CLI wording exactly: iOS AgentPro in the row above the install
button, macOS and Android beside the review action, and Control UI on the row
that explains why review is refused for these sources.

Covered by a wire assertion that the state reaches clients for an external
source and stays absent for registry rows, plus decode-and-label tests on the
shared Swift kit and the Android parser, and a Control UI render assertion.

* fix(ui): size the ClawHub detail dialog to a refusal message

Refusing detail for an external source made an error-only dialog reachable.
The shared preview panel reserves a tall reader height for skill documents, so
a two-line refusal rendered in a mostly empty dialog and read as broken rather
than deliberate. Found by inspecting the review captures.

* revert(ui,apps): drop the ClawHub trust label layer

Maintainer product decision: skills.sh runs its own scanners, so OpenClaw does
not add a second alert layer in the apps. Removes the label from Control UI,
iOS, macOS and Android, and drops the trustState wire field that nothing would
render. The CLI keeps its existing label; changing that is a separate call.

Publisher identity, the fail-closed detail refusal, and the message-only dialog
are unchanged. Splits the oversized skills view test file to satisfy max-lines
without a suppression.

* test(ui): fix ClawHub skill fixture checks

* chore(plugin-sdk): refresh API baseline

---------

Co-authored-by: Patrick Erichsen <patrick.a.erichsen@gmail.com>
2026-08-12 20:13:15 -07:00
..

@openclaw/gateway-protocol

Typed schemas, inferred TypeScript types, and runtime validators for the OpenClaw Gateway WebSocket protocol.

The current wire protocol is version 4. General clients must use v4; authenticated node clients and lightweight probes may use the N-1 window during rolling upgrades. See the Gateway protocol specification for transport, authentication, roles, scopes, and complete frame examples.

Versioning

Package versions follow the OpenClaw calendar release train: YYYY.M.PATCH, with the same prerelease suffix when applicable. A package version therefore identifies the OpenClaw source release that produced the schemas; it is not the wire protocol number.

The wire protocol integer is versioned separately. Its current value is exported as PROTOCOL_VERSION from @openclaw/gateway-protocol/version. Gateway protocol changes are additive first. An incompatible wire change requires an explicit protocol-version decision and coordinated client follow-through. See CHANGELOG.md for the wire and schema history.

Install

npm install @openclaw/gateway-protocol

Entry points

  • @openclaw/gateway-protocol exports runtime validators, selected schemas, error formatting, and their TypeScript types. This is the main TypeBox-backed entry.
  • @openclaw/gateway-protocol/schema exports the TypeBox schema graph, including the ProtocolSchemas registry used by generators.
  • @openclaw/gateway-protocol/frame-guards exports dependency-free structural guards for gateway event and response envelopes.
  • @openclaw/gateway-protocol/client-info exports client ID, mode, and capability registries plus normalization helpers.
  • @openclaw/gateway-protocol/connect-error-details exports structured connect error readers and recovery metadata.
  • @openclaw/gateway-protocol/gateway-error-details exports helpers for reading structured details from general gateway errors.
  • @openclaw/gateway-protocol/startup-unavailable exports startup retry constants and helpers.
  • @openclaw/gateway-protocol/version exports the current and minimum accepted protocol versions.

The frame-guards, client-info, connect-error-details, gateway-error-details, startup-unavailable, and version entry points are TypeBox-free. Prefer them when a browser bundle only needs envelope dispatch, handshake constants, or reconnect policy. This also avoids runtime compilation in CSP-sensitive consumers. The root and schema entry points provide the full validation surface and depend on TypeBox.

Validate an inbound frame

The compiled validators are callable type guards. Their errors property contains the most recent validation errors.

import { formatValidationErrors, validateRequestFrame } from "@openclaw/gateway-protocol";

const frame: unknown = JSON.parse(inboundText);

if (!validateRequestFrame(frame)) {
  throw new Error(formatValidationErrors(validateRequestFrame.errors));
}

console.log(frame.id, frame.method);

validateRequestFrame validates the request envelope. Dispatch code must also use the validator for the selected method's params; the root entry point exports those validators as validate*Params functions.

Guard an event without TypeBox

Use the lightweight guards when code only needs safe frame discrimination. They check dispatch-critical envelope fields and intentionally allow additive payload fields.

import { isGatewayEventFrame } from "@openclaw/gateway-protocol/frame-guards";

const frame: unknown = JSON.parse(inboundText);

if (isGatewayEventFrame(frame)) {
  console.log(frame.event, frame.seq);
}

Build handshake version and capability fields

Protocol levels and client capabilities live in TypeBox-free entry points.

import { GATEWAY_CLIENT_CAPS } from "@openclaw/gateway-protocol/client-info";
import { MIN_CLIENT_PROTOCOL_VERSION, PROTOCOL_VERSION } from "@openclaw/gateway-protocol/version";

const handshake = {
  minProtocol: MIN_CLIENT_PROTOCOL_VERSION,
  maxProtocol: PROTOCOL_VERSION,
  caps: [GATEWAY_CLIENT_CAPS.TOOL_EVENTS],
};

Nodes and probes use MIN_NODE_PROTOCOL_VERSION and MIN_PROBE_PROTOCOL_VERSION, respectively. A capability advertises client support; it does not grant authorization.

Contract notes

Session identifiers

Several identifier names coexist because they identify different things:

  • key is the established logical session selector used by most sessions.* CRUD, send, subscription, patch, reset, delete, compaction, and usage methods. A key can be canonicalized or resolved within an agent's session store.
  • sessionKey names the same logical routing identity where the contract needs to make that meaning explicit. chat.*, session file and diff APIs, transcript branch/rewind/fork APIs, agent events, and channel delivery payloads use this spelling.
  • sessionId is the opaque stored transcript or runtime instance ID. Session results may return it beside a key. Talk, terminal, worker, and selected channel protocols also use sessionId for their own concrete session instances; do not substitute a logical session key there.

Follow each method schema rather than converting fields based on their spelling. sessions.resolve is the explicit bridge when a caller has a key, raw session ID, label, Control UI short ID, or parent/agent scope.

Intentionally open fields

The schema graph is strict by default, but roughly 60 fields intentionally use Type.Unknown() passthroughs. The main clusters are transport-owned channel payloads, logs-chat message and attachment passthrough, worker and node tool arguments/results, and the dynamic config.schema response. Frame params, payload, and error details are also open at the envelope layer because the selected method, event, or error code owns their concrete shape.

Do not treat these fields as validated domain objects. Narrow them at their owner boundary before reading nested values.

Machine-readable schema

protocol.schema.json ships in the npm tarball as the generated machine-readable contract. It contains the frame union, named schema definitions, and core method metadata. It is generated during prepack and is not committed to the repository.

Method discovery

The hello-ok.features.methods list is conservative discovery, not a complete enumeration of every callable method. It reflects the methods the connected Gateway intentionally advertises. Core-internal, role-specific, plugin-provided, or otherwise non-advertised methods can have valid schemas without appearing in that list. Clients should use discovery to enable optional UI, not to reject an otherwise documented method contract.