Landing gate research (codex sweep of 77 external channel packages — catalogs, npm, ClawHub; full file:line evidence archived in the maintainer notes): 13 packages use the surface this PR deletes. Flat/scalar streaming config ships in 11 (notably @srinathh/openclaw-channel-twilio-whatsapp carries flat chunkMode in its own plugin.json defaults — operator-side doctor cannot fix that; only a plugin update can), and resolveGroupIntroHint is registered by @soimy/dingtalk (4.3k downloads), @tencent-connect/openclaw-qqbot, @cxyhhhhh/openclaw-qqbot, @nevis/talk (type-level break only; core never read the hint). The deprecated channel-streaming subpath (used by @inline-openclaw/inline and @roamhq/openclaw-roam) is NOT deleted here. Tag math: the warnings first shipped in v2026.7.2-beta.1; v2026.7.1 stable predates them. Earliest policy-compliant landing: after the 2026.7.2 STABLE tag (which must also pick up the intro-hint deprecation from #107252), into the 2026.7.3 train — realistically 2026-07-16/17 at current cadence, ideally after notifying the affected package owners.

Deprecation umbrella issue with the affected-package list, migration snippet, and train timeline: #107860

Codex review: needs real behavior proof before merge. Reviewed July 15, 2026, 12:23 AM ET / 04:23 UTC.

Summary
Removes flat and scalar channel-streaming compatibility reads, removes the deprecated group intro-hint adapter and WhatsApp export, and rejects legacy Matrix account streaming scalars with a doctor pointer.

PR surface: Source -185, Tests -227, Docs -2, Generated 0. Total -414 across 29 files.

Reproducibility: not applicable. This PR intentionally removes deprecated compatibility behavior rather than fixing a reported current-main bug. The affected upgrade paths are clearly identified from source and the external-package audit.

Review metrics: 3 noteworthy metrics.

• Deprecated config spellings: 5 compatibility forms removed. Four flat delivery keys plus scalar streaming are still used by external packages and can change upgrade behavior.
• Public plugin surfaces: 1 adapter field and 1 export removed. These removals create compile-time compatibility breaks for third-party plugin code.
• Known external consumers: 13 packages affected. The ecosystem sweep establishes concrete blast radius beyond hypothetical compatibility concerns.

Stored data model
Persistent data-model change detected: migration/backfill/repair: docs/concepts/streaming.md, migration/backfill/repair: extensions/matrix/src/config-schema.test.ts, migration/backfill/repair: extensions/matrix/src/config-schema.ts, migration/backfill/repair: src/commands/doctor/shared/deprecation-compat.ts, serialized state: extensions/matrix/src/config-schema.test.ts. Confirm migration or upgrade compatibility proof before merge.

Root-cause cluster
Relationship: fixed_by_candidate
Canonical: #107860
Summary: The open deprecation notice issue is the canonical release and ecosystem-migration tracker; this PR is its candidate removal implementation.

Members:

• canonical: Deprecation notice: flat/scalar streaming config keys and resolveGroupIntroHint removal (2026.7.3 train) #107860 - It tracks the affected packages, migration instructions, stable warning window, and intended removal train.
• partial_overlap: refactor(channels): deprecate the never-consumed group intro hint adapter #107252 - The merged PR introduced the group intro-hint deprecation that this branch proposes to complete.

Proposal only: this assessment does not dispatch repair, suppress jobs, mutate sibling items, close, or merge anything.

Merge readiness
Overall: 🦪 silver shellfish
Proof: 🦪 silver shellfish
Patch quality: 🐚 platinum hermit
Result: blocked until real behavior proof from a real setup is added.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

• [P1] Add redacted real-setup upgrade proof covering legacy failure, doctor migration, restart, and preserved nested behavior.
• After v2026.7.2 stable, record explicit approval for the public API removal train and affected-package migration plan.
• Rebase onto current main and refresh the review against the conflict-resolved head.

Proof guidance:

• [P1] Needs real behavior proof before merge: Automated Testbox suites are supplemental only; add a redacted terminal or live upgrade transcript showing legacy rejection, openclaw doctor --fix, restart, and nested behavior, then update the PR body or ask a maintainer for @clawsweeper re-review. Redact keys, phone numbers, IPs, endpoints, and other private data.

Risk before merge

• [P1] Landing before v2026.7.2 stable would break known external consumers without any stable-release warning window; one affected plugin embeds legacy defaults that doctor cannot repair.
• [P1] The branch removes a public Plugin SDK field and export despite scoped policy reserving breaking removals for major-version work, so the intended 2026.7.3 exception requires explicit owner approval.
• [P1] Existing legacy configurations can silently fall back to different streaming behavior, while Matrix account scalar configurations can block startup until doctor or plugin updates are applied.
• [P1] The current head is draft and conflict-marked against newer main, so the exact rebased merge result still needs review.

Maintainer options:

1. Finish the deprecation window (recommended)
   Ship v2026.7.2 stable, migrate or notify known consumers, rebase the branch, and add live upgrade proof before reconsidering merge.
2. Preserve the public compatibility layer
   Keep the deprecated reads and adapter types until an explicitly approved major-version removal window.
3. Own the breaking train
   A release owner may accept the 2026.7.3 break with explicit compatibility approval and clear upgrade communication.

Next step before merge

• [P1] Hold the draft until v2026.7.2 stable ships, then have the compatibility owner confirm the breaking-release exception, affected-package handling, real upgrade proof, and rebased merge result.

Maintainer decision needed

• Question: After v2026.7.2 stable ships the warnings, may the 2026.7.3 train intentionally remove these public compatibility surfaces, or must they remain until a major-version boundary?
• Rationale: The code cleanup matches the canonical-config direction, but repository Plugin SDK policy, 13 known external consumers, and the absence of a stable warning release make the permissible removal train a maintainer-owned compatibility decision.
• Likely owner: steipete — Their merged deprecation work, ecosystem audit, tracking issue, and current branch provide the strongest evidence of ownership over the removal plan.
• Options:
  • Wait, migrate, and approve (recommended): After v2026.7.2 stable, notify or update affected packages, obtain explicit approval for the breaking train, rebase, and require live upgrade proof before merge.
  • Keep compatibility until major: Retain the deprecated resolver and adapter surfaces until the next explicitly designated major-version boundary.
  • Accept the 2026.7.3 break: Land after the stable tag while explicitly accepting plugin breakage and documenting required operator and package-owner action.

Security
Cleared: The diff changes compatibility resolution, validation, tests, docs, and generated API hashes without adding dependencies, executable downloads, workflow permissions, secret access, or package-resolution changes.

Superseded by #113533, which carries the maintainer-approved v2026.7.2 compatibility decision, current reviewed scope, exact-head proof, and final landing validation.

Landed on main as d2b6573.