docs: propose OpenClaw-owned Rust node runtime - #54
giodl73-repo wants to merge 36 commits into
Conversation
|
Codex review: needs maintainer review before merge. Reviewed September 16, 2026, 11:15 PM ET / September 17, 2026, 03:15 UTC (Revision 61). ClawSweeper reviewWhat this changesAdds a Rust node-runtime ownership RFC with companion contracts, adoption gates, draft connection-control fixtures, an architecture diagram, and implementation evidence. Merge readiness✅ Ready for maintainer review Keep open for landing: this distinct RFC is absent from the target default branch. GitHub-verified write access supports the recorded bounded acceptance, resolving the previous maintainer-decision blocker. No actionable patch defect remains. Priority: P3 Review scores
Verification
How this fits togetherThe proposed Rust runtime connects native applications to OpenClaw’s Gateway and dispatches authorized node commands. Gateway policy remains authoritative, while embedding products retain credentials, native tools, packaging, and deployment. flowchart LR
A[Authorized controllers] --> B[OpenClaw Gateway]
B --> C[Rust Gateway client]
C --> D[Bounded node runtime]
D --> E[Product admission and adapters]
E --> F[Native tools]
F --> D
D --> B
Before mergeNone. Agent review detailsSecurityNone. Review metrics
Technical reviewBest possible solution: Record the bounded monorepo ownership decision while preserving independent conformance, native-adoption, and supported-release gates. Do we have a high-confidence way to reproduce the issue? Not applicable: this is a design-record PR, and the previously reported message/frame documentation mismatch is corrected. Is this the best way to solve the issue? Yes: the bounded ownership record matches the landed foundation and explicitly separates unimplemented connection control and product rollout from RFC acceptance. AGENTS.md: not found in the target repository. Codex review notes: model internal, reasoning medium; reviewed against 967d9aac7472. LabelsLabel changes:
Label justifications:
EvidenceWhat I checked:
Likely related people:
Rating scale
Overall follows the weaker of proof and patch quality. Workflow
HistoryReview history (60 earlier review cycles; latest 8 shown)
|
|
@clawsweeper re-review |
|
🦞🧹 I asked ClawSweeper to review this item again. |
|
🦞🧹 I asked ClawSweeper to review this item again. Re-review progress:
|
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 2c44506f-c357-47e5-a867-91775df2093d
|
@clawsweeper re-review The RFC evidence inventory is refreshed to the final follow-up heads:
Please review exact RFC head |
|
🦞👀 Re-review progress:
|
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 2c44506f-c357-47e5-a867-91775df2093d
|
@clawsweeper re-review The RFC evidence inventory is refreshed to the final follow-up heads:
Please review exact RFC head |
|
🦞🧹 I asked ClawSweeper to review this item again. Re-review progress:
|
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 2c44506f-c357-47e5-a867-91775df2093d
|
@clawsweeper re-review The RFC evidence inventory is refreshed to the final follow-up heads:
Please review exact RFC head |
|
🦞🧹 I asked ClawSweeper to review this item again. Re-review progress:
|
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 2c44506f-c357-47e5-a867-91775df2093d
|
@clawsweeper re-review |
|
🦞🧹 I asked ClawSweeper to review this item again. Re-review progress:
|
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 2c44506f-c357-47e5-a867-91775df2093d
|
@clawsweeper re-review |
|
🦞🧹 I asked ClawSweeper to review this item again. Re-review progress:
|
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 2c44506f-c357-47e5-a867-91775df2093d
|
@clawsweeper re-review |
|
🦞🧹 I asked ClawSweeper to review this item again. Re-review progress:
|
|
@clawsweeper approve Maintainer decision: accept the bounded, unpublished two-crate Rust foundation described by this RFC. This is an explicit design-only proof override for the deferred and currently unimplemented Gateway connection-control exchange. It does not approve that exchange for adoption without runtime proof, publish crates, ship an official binary, authorize broader |
|
🦞👀 Command router queued. I will update this comment with the next step. |
|
@clawsweeper re-review |
|
🦞🧹 I asked ClawSweeper to review this item again. Re-review progress:
|
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 2c44506f-c357-47e5-a867-91775df2093d
|
@clawsweeper re-review |
|
🦞🧹 I asked ClawSweeper to review this item again. Re-review progress:
|
|
🦞👀 Pull request received. I will update this pull request when review starts. ClawSweeper review completeClawSweeper finished reviewing this revision. The review result is being finalized. |
|
@clawsweeper re-review |
|
🦞🧹 I asked ClawSweeper to review this item again. Re-review progress:
|
|
@steipete RFC #54 is accepted and ready at exact head ClawSweeper's final review is 5/6 with no findings or remaining pre-merge work. The accepted scope is the bounded unpublished two-crate foundation; the design-only proof override is limited to the deferred Gateway connection-control exchange, and all adoption and release gates remain intact. The repository's |
What problem this solves
Native OpenClaw products need a reusable Rust Gateway client and bounded node
runtime without duplicating OpenClaw protocol, lifecycle, policy, and security
behavior in each product repository.
Decision
OpenClaw owns two unpublished workspace crates:
openclaw-gateway-clientfor role-neutral transport and session mechanics;openclaw-node-hostfor the bounded node role, lifecycle, runtime, andheadless proof host.
TypeScript and the Gateway remain authoritative for protocol, pairing,
approvals, policy, command semantics, and cancellation delivery. Native
products retain credentials, TLS trust, signing, native admission and tools,
packaging, proxy routing, rollout, and support policy.
RFC acceptance does not publish crates, promise stable Rust APIs, ship an
official binary, require a sidecar, or authorize broader
system.*, worker,workspace-transfer, plugin, PTY, MCP, or skills parity.
The maintainer acceptance decision applies to the bounded, already-proven
two-crate incubation contract. It is also the design-only proof override for
the explicitly deferred Gateway connection-control exchange; that future
contract still requires implementation and runtime proof before adoption.
Landed implementation
The bounded foundation has now merged into
openclaw/openclawin dependencyorder:
client and bounded host, merged as
f9a7f104c22a.lifecycle and native ownership hooks, merged as
29069179def6.sidecar bridge and separate-process proof, merged as
ce4f1d711bbc.Current follow-ups
3083030ee1amakes the bounded TypeScript/Rust lifecycle fixture bilateraland adds a permanent path-scoped conformance workflow.
a97e9bdba67isolates reusable native signing, hostname-bound TLS trust,admission, request-lifetime, cancellation, and non-starvable keepalive seams.
#149725 at
f0ada44a58ais now a focused helper, Swift adapter, packaging/signing,proxy-safe routing, and native-probe stack on #150344.
These follow-ups strengthen conformance and adoption evidence but do not expand
the bounded RFC decision. The macOS adopter remains draft until its prerequisite
lands and exact-head package, signing, install, upgrade, rollback, TLS, proxy,
lifecycle, saturation, and performance evidence is complete.
Evidence
Rust/Tauri proof, strict static gates, and clean independent review before
landing.
separate-process proofs.
contract/registry tests, 639 workflow-guard tests, strict Rust static gates,
formatting, and its exact-head Node Runtime Conformance workflow.
Periphery, macOS Swift tests, and iOS smoke.
Architecture sketch