Best possible solution:

Land the narrow release-packager compatibility fix once exact-head CI is acceptable, while preserving hard failures for malformed AI-workspace package sources.

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

Yes from source inspection: current main unconditionally rejects package sources without dependencies.@openclaw/ai before npm pack, so a pre-AI-workspace source ref without that dependency hits the failure path. I did not run a historical ref because this review was read-only.

Is this the best way to solve the issue?

Yes, this is the narrowest maintainable fix I found: keep the compatibility decision inside the package helper and preserve strict errors for inconsistent AI-workspace sources. Moving the workaround into release-candidate resolution would duplicate packaging policy across callers.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 1a37cafd63b6.

Label changes:

• add P2: This is a narrow release-packaging regression fix with limited blast radius and no evidence of an active user-facing outage.
• add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: This is a maintainer-authored PR with the maintainer label, so the external-contributor real-behavior-proof gate does not apply.

Label justifications:

• P2: This is a narrow release-packaging regression fix with limited blast radius and no evidence of an active user-facing outage.
• rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: This is a maintainer-authored PR with the maintainer label, so the external-contributor real-behavior-proof gate does not apply.

PR surface:

Tests +62, Other +18. Total +80 across 2 files.

What I checked:

• Current main failure path: Current main reads the source package.json and rejects any package source whose dependencies do not declare @openclaw/ai before the npm pack step can run. (scripts/package-openclaw-for-docker.mjs:456, 1a37cafd63b6)
• PR compatibility guard: The PR head checks for packages/ai/package.json, returns a no-op cleanup only when both the workspace and dependency are absent, and still errors when the dependency exists without the workspace. (scripts/package-openclaw-for-docker.mjs:438, b33cd8c32587)
• Caller boundary: The Docker packer calls prepareBundledAiRuntimePackage before npm pack and then validates the resulting tarball, so the compatibility branch is localized to release-package preparation. (scripts/package-openclaw-for-docker.mjs:570, 1a37cafd63b6)
• Tarball validation contract: The tarball checker only enforces bundled @openclaw/ai when the packed package actually declares that dependency, which matches the PR's legacy-no-dependency branch. (scripts/check-openclaw-package-tarball.mjs:126, 1a37cafd63b6)
• Regression coverage: The PR adds tests for unchanged pre-AI-workspace sources and for incomplete AI workspace sources, while updating the existing positive AI-runtime bundling fixture to include packages/ai/package.json. (test/e2e/qa-lab/runtime/package-openclaw-for-docker.e2e.test.ts:237, b33cd8c32587)
• Formatting sanity: The exact PR diff for the two touched files has no whitespace errors. (b33cd8c32587)

Likely related people:

• vincentkoc: Recent current-main history touches the same release package and cross-OS candidate packaging surface, including package tarball runtime checks and pnpm candidate packing. (role: recent area contributor; confidence: high; commits: 63257a2b0eec, 0cd2bc474fa8, a1aaa1c6debb; files: scripts/package-openclaw-for-docker.mjs, scripts/check-openclaw-package-tarball.mjs, scripts/openclaw-cross-os-release-checks.ts)
• steipete: The @openclaw/ai workspace split in refactor: extract reusable AI runtime package #99059 created the packaging invariant this PR now makes backward-compatible for older refs. (role: adjacent feature owner; confidence: medium; commits: 42aa617a9f86, 43625f733916, 1833c0d77180; files: packages/ai/package.json, scripts/package-openclaw-for-docker.mjs, test/e2e/qa-lab/runtime/package-openclaw-for-docker.e2e.test.ts)

• 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
• 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
• 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
• 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
• 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
• 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
• 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

• ClawSweeper keeps one durable marker-backed review comment per issue or PR.
• Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
• A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
• PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
• Maintainers can also comment @clawsweeper review to request a fresh review only.
• Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
• Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
• Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.