RFC 0012: Gateway-Independent OpenClaw Core

[fuller-stack-dev]

Summary

• Propose a Host-owned OpenClaw core that keeps conversations, sessions, cron, heartbeat, tasks, and durable delivery intent running independently of channel connectivity and Control Server restarts.
• Preserve the existing openclaw gateway command, gateway.* config, protocol, UI/TUI/CLI behavior, and external plugins through a Compatibility Supervisor and automatic legacy co-located placement.
• Make configuration reload and restart behavior owner-scoped, then roll out the logical and physical split through proof-gated phases.
• Include a detailed implementation-plan sidecar covering ordered work packages, plugin-developer compatibility, rollout gates, risks, and fault/upgrade proof.

Visual Proposal

• Pagedrop: Gateway-Independent OpenClaw Core implementation proposal

Verification

• git diff --check origin/main...HEAD
• RFC frontmatter, required heading order, and one-paragraph summary validator
• RFC-relative link validator
• Current-source survey against openclaw/openclaw main at 127e174c9e4de36f2eccb96300bb3efa69ae9a32

Process Follow-up

• Create the required maintainer-discussion thread for the RFC.

[clawsweeper]

Codex review: needs changes before merge.

What this changes

This PR proposes separating core agent work from channel and control-server lifecycles through explicit owners and a phased implementation plan.

Example: Change gateway.port from 18789 to 18790 during an active agent turn.

• Before: The RFC describes a shared Gateway restart that can interrupt core work.
• After: The proposed architecture restarts only the Control Server while the Host continues the turn; this PR adds no runtime implementation.

Review scores

Measure	Result	What it means
Overall readiness	🐚 platinum hermit (4/6)	A coherent documentation proposal with explicit ownership, compatibility, rollout, and validation boundaries; no discrete patch defect was found.
Proof confidence	🌊 off-meta tidepool	Not applicable: This documentation-only PR specifies future runtime behavior and proof requirements without changing a production owner.
Patch quality	🐚 platinum hermit (4/6)	No actionable review findings were identified.

Product

Kind: Docs · Worth it: Yes
User problem: The proposal targets interruptions to ongoing agent work when transport or control-server lifecycles restart.
Reason: The member-authored direction has concrete reliability value and preserves existing contracts; repository-required RFC acceptance remains a lifecycle step rather than a second product-direction request.

Merge readiness

⛔ Needs changes before merge - 1 item remains

Keep open: this member-authored RFC remains distinct from main and provides a coherent compatibility-preserving roadmap; its draft lifecycle still needs completion.

Priority: P2
Reviewed head: ffa5250ddcb49ff62bd830e7fda66a30dfa90ef4

Before merge

• Complete next step - Create and link the required maintainer-discussion thread; if accepted, create the implementation issue, update status and issue frontmatter, and mark the PR ready before merge.

Findings

None.

Agent review details

How this fits together

This repository records architectural proposals; the proposed runtime would route client requests and channel ingress into Host-owned core services, then send delivery intent through channel connectors.

flowchart LR
  A[Existing clients] --> B[Control Server]
  C[Compatibility Supervisor] --> B
  C --> D[Host core]
  C --> E[Channel connectors]
  B --> D
  E --> D
  D --> F[Sessions and scheduled work]
  D --> G[Durable delivery intent]
  G --> E

Loading

Technical review

Best possible solution:

Complete the RFC lifecycle, then implement the ownership split incrementally under its compatibility and fault-proof gates.

Do we have a high-confidence way to reproduce the issue?

This is a design proposal, not a bug-fix implementation; its historical core-source survey does not establish a current-main reproduction.

Is this the best way to solve the issue?

Logical isolation before process separation is a coherent approach, with one core implementation and explicit compatibility gates; future implementation PRs must establish the promised behavior.

AGENTS.md: not found in the target repository.

Codex review notes: model internal, reasoning medium; reviewed against 967d9aac7472.

Security

None.

Evidence

What I checked:

• Verified introduction: The pinned introduction adds only the RFC and its implementation-plan sidecar: 919 Markdown lines, with no executable runtime change. (rfcs/0012-gateway-independent-core.md:16, ffa5250ddcb4)
• RFC lifecycle: Main requires a maintainer-discussion thread, acceptance, an implementation issue, and accepted frontmatter before merge; the proposal currently has draft status and a blank issue. (README.md:76, 967d9aac7472)
• Current PR state: GitHub confirms MEMBER authorship, the reviewed head, and an open draft PR; its body still lists creating the discussion thread as follow-up, and no human review or discussion link is recorded in the inspected GitHub conversation. (ffa5250ddcb4)
• Distinct proposal: The fetched main tree contains neither proposed file. Related proposals RFC 0013 addendum: application-consistent recovery for scale-to-zero #46, docs: propose guest foreground work and restart notices #74, and docs(rfc): propose external-supervisor operator guidance #75 address recovery handoff, guest interruption notices, and external-supervisor guidance respectively, rather than this four-owner core separation. (967d9aac7472)
• Compatibility and implementation gates: The RFC retains public commands, configuration, protocol, and legacy plugin placement; its sidecar requires fault, security, upgrade, and rollback proof for future implementations rather than claiming those implementations exist. (rfcs/0012/implementation-plan.md:413, ffa5250ddcb4)
• RFC-process history: GitHub commit patches verify kevinslin's lifecycle contribution and RomneyDa's sidecar-layout contribution in commit 3aa7d72. Local README blame failed on an unavailable historical blob; the commit API recovered the relevant patches, but runtime feature ownership remains unestablished. (README.md:76, e366ea9825a4)

Review metrics

Metric	Value	Why it matters
Proposal scope	Documentation +919 lines; production +0; tests +0	The sidecar supplies phased work packages and implementation gates without changing shipped behavior.

Labels

Label changes:

• add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• add status: ⏳ waiting on author: ClawSweeper has contributor-facing work open and is waiting for author action.
• remove status: 👀 ready for maintainer look: Current PR status label is status: ⏳ waiting on author.
• remove rating: 🦞 diamond lobster: Current PR rating is rating: 🐚 platinum hermit, so this older rating label is no longer current.

Label justifications:

• P2: This is substantive architectural work for normal consideration, with no immediate runtime regression introduced.
• rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• status: ⏳ waiting on author: ClawSweeper has contributor-facing work open and is waiting for author action.

Rating scale

6/6 🦀 challenger crab · 5/6 🦞 diamond lobster · 4/6 🐚 platinum hermit · 3/6 🦐 gold shrimp · 2/6 🦪 silver shellfish · 1/6 🧂 unranked krab. Overall follows the weaker of proof and patch quality; ✨ marks media proof (a screenshot, video, or linked artifact) that directly shows the changed behavior.

Workflow

ClawSweeper edits this one comment on every review. Comment @clawsweeper re-review for a fresh review only. Repair and merge execution lanes are retired.

History

Review history (59 earlier review cycles; latest 8 shown)

• reviewed 2026-08-31T09:08:37.073Z sha ffa5250 :: needs maintainer review before merge. :: none
• reviewed 2026-09-01T04:04:15.673Z sha ffa5250 :: needs maintainer review before merge. :: none
• reviewed 2026-09-02T20:13:08.231Z sha ffa5250 :: blocked before merge. :: none
• reviewed 2026-09-02T23:54:07.140Z sha ffa5250 :: blocked before merge. :: none
• reviewed 2026-09-03T04:53:31.062Z sha ffa5250 :: blocked before merge. :: none
• reviewed 2026-09-05T06:55:07.184Z sha ffa5250 :: blocked before merge. :: none
• reviewed 2026-10-01T19:44:43.493Z sha ffa5250 :: blocked before merge. :: none
• reviewed 2026-10-07T21:47:16.797Z sha ffa5250 :: blocked before merge. :: none

Reviewed October 12, 2026, 12:37 PM ET / 16:37 UTC (Revision 60).