Repository navigation
Description
Summary
The running GHCR/Docker OpenClaw artifact for 2026.5.28 emits a CLI config schema that rejects channels.discord.streaming.progress.commentary, even though the source/tag and some installed runtime files indicate that Discord progress commentary support should be present.
This is a packaging/generated-artifact report, not a request to re-implement #85200.
The previous source-level issue was closed as already implemented: #88779. ClawSweeper specifically suggested treating any remaining container-only mismatch as a fresh packaging report with image evidence, so this issue focuses on the running container artifact.
Environment
- Deployment source: GHCR OpenClaw Docker image (
ghcr.io/openclaw/...) - Runtime path: containerized OpenClaw,
/app/openclaw.mjs - Running container hostname:
8f713a0550a4 - Running Docker container id from cgroup:
8f713a0550a46406bb9be63d98a3d112de55365dbe8c7815634280de5431a610 - Docker socket: not mounted in this container
- Portainer API key: not available inside this container
- GHCR manifest/digest lookup from inside the container: unauthenticated registry requests returned
401
Because there is no Docker socket and no Portainer API key inside the container, I cannot include the image digest from this runtime. But the container was pulled from GHCR and the package/runtime version below is from inside that running container.
Running version proof
node /app/openclaw.mjs --version
node -e "const p=require('/app/package.json'); console.log(JSON.stringify({packageName:p.name,packageVersion:p.version},null,2))"
hostname
cat /proc/self/cgroup | head -20Output:
OpenClaw 2026.5.28
{
"packageName": "openclaw",
"packageVersion": "2026.5.28"
}
8f713a0550a4
9:blkio:/docker/8f713a0550a46406bb9be63d98a3d112de55365dbe8c7815634280de5431a610
8:devices:/docker/8f713a0550a46406bb9be63d98a3d112de55365dbe8c7815634280de5431a610
7:freezer:/docker/8f713a0550a46406bb9be63d98a3d112de55365dbe8c7815634280de5431a610
6:memory:/docker/8f713a0550a46406bb9be63d98a3d112de55365dbe8c7815634280de5431a610
5:cpu:/docker/8f713a0550a46406bb9be63d98a3d112de55365dbe8c7815634280de5431a610
4:cpuset:/docker/8f713a0550a46406bb9be63d98a3d112de55365dbe8c7815634280de5431a610
3:cpuacct:/docker/8f713a0550a46406bb9be63d98a3d112de55365dbe8c7815634280de5431a610
2:name=synomonitor:/system.slice/pkg-ContainerManager-dockerd.service
1:name=systemd:/docker/8f713a0550a46406bb9be63d98a3d112de55365dbe8c7815634280de5431a610
Reproduction
Inside that running container:
node /app/openclaw.mjs config schema > /tmp/openclaw-schema-current.json
node -e "const s=require('/tmp/openclaw-schema-current.json'); const props=s.properties.channels.properties.discord.properties.streaming.properties.progress.properties; console.log('hasCommentary=' + Object.prototype.hasOwnProperty.call(props,'commentary')); console.log(Object.keys(props).sort().join(','));"Output:
hasCommentary=false
commandText,label,labels,maxLineChars,maxLines,render,toolProgress
Attempting to validate the documented/implemented setting fails:
node /app/openclaw.mjs config set channels.discord.streaming.progress.commentary true --strict-json --dry-runOutput:
Error: Dry run failed: config schema validation failed.
- - channels.discord.streaming.progress: invalid config: must not have additional properties: "commentary"
Why this looks like an artifact mismatch
The same installed /app/dist tree contains runtime/UI-hint references to the field:
grep -RIn --exclude-dir=node_modules "streaming.progress.commentary\\|Discord Progress Commentary\\|resolveChannelStreamingProgressCommentary\\|DiscordStreamingProgressSchema" /app/dist 2>/dev/null | head -80Relevant output from the same running container:
/app/dist/bundled-channel-config-schema-B8zs_onv.js:120:const DiscordStreamingProgressSchema = ChannelStreamingProgressSchema.extend({ commentary: boolean().optional() }).strict();
/app/dist/bundled-channel-config-schema-B8zs_onv.js:130:const DiscordPreviewStreamingConfigSchema = ChannelPreviewStreamingConfigSchema.extend({ progress: DiscordStreamingProgressSchema.optional() }).strict();
/app/dist/config-schema-BeUb0EJF.js:95: "streaming.progress.commentary": {
/app/dist/config-schema-BeUb0EJF.js:96: label: "Discord Progress Commentary",
/app/dist/message-handler.process-CP-OZdtT.js:610: const commentaryProgressEnabled = Boolean(draftStream) && resolveChannelStreamingProgressCommentary(params.discordConfig);
/app/dist/streaming-Bv78t77e.js:412:function resolveChannelStreamingProgressCommentary(entry, defaultValue = false) {
So the running artifact appears internally inconsistent:
- Runtime/UI-hint files include
streaming.progress.commentary. - The CLI schema emitted by
node /app/openclaw.mjs config schemaomitscommentary. - Strict validated config writes reject the field.
Expected behavior
For a 2026.5.28 GHCR image that includes the #85200 runtime/docs/schema work, this should validate:
{
"channels": {
"discord": {
"streaming": {
"mode": "progress",
"progress": {
"commentary": true
}
}
}
}
}At minimum, the generated CLI schema in the shipped Docker image should agree with the runtime resolver and UI hints in the same image.
Actual behavior
The GHCR-pulled running container reports OpenClaw 2026.5.28 / openclaw@2026.5.28, but its live CLI schema rejects channels.discord.streaming.progress.commentary.
Related context
- Discord progress drafts should support assistant commentary events #83307: original feature request, "Discord progress drafts should support assistant commentary events"
- Show assistant commentary in Discord progress drafts #85200: merged implementation, "Show assistant commentary in Discord progress drafts"
- Discord progress commentary config is present in runtime but rejected by schema #88779: earlier issue closed as already implemented in source; follow-up comments include the same live container evidence
Suggested investigation
Please check whether the GHCR Docker image published for 2026.5.28 includes a stale generated root config schema or mixed generated artifacts. A useful regression would be an image/package-level smoke test that runs:
node /app/openclaw.mjs config schemaand asserts that channels.discord.streaming.progress.properties.commentary exists when the Discord runtime bundle includes resolveChannelStreamingProgressCommentary.
Metadata
Metadata
Assignees
Labels
Type
Fields
Priority
Activity
clawsweeper commentedon May 31, 2026
Thanks for the report. I gave this a fresh shell check against current
main, and I could not reproduce it anymore.Close: the exact GHCR 2026.5.28 Docker path now emits and validates the Discord progress commentary config key, and the current latest/2026.6.8 image passes the same checks.
Review details
Best possible solution:
Keep the current GHCR/source behavior; if a future container shows this again, open a new digest-specific artifact report with fresh schema and strict dry-run output.
Do we have a high-confidence way to reproduce the issue?
No. The exact current GHCR
2026.5.28amd64 image now emits thecommentaryschema property and accepts the strict dry-run config write, so the reported failure did not reproduce.Is this the best way to solve the issue?
Yes. Closing as no longer reproducible is the narrowest current outcome because both the historical tag and the latest image pass the reported schema/config checks.
Security review:
Security review: No patch is under review, so there is no diff-level security review to perform.
AGENTS.md: found and applied where relevant.
What I checked:
ghcr.io/openclaw/openclaw:2026.5.28now printsOpenClaw 2026.5.28,hasCommentary=true, and progress keys includingcommentary.Dry run successful: 1 update(s) validated against $OPENCLAW_HOME/.openclaw/openclaw.json..2026.5.28tag resolves to index digestsha256:dbad4655fb87aa8d8662323daf9f46f4721504324e22c9f99d84849e609f5b2c, and the linux/amd64 child issha256:4e23676146b60da884d4a6ee050086cad7478c6041853f56eab9db0aad002921, matching the reporter's digest evidence.latestresolves to OpenClaw 2026.6.8 and also emitshasCommentary=trueand accepts the strict dry-run config write.commentary: z.boolean().optional()in the shared channel progress streaming schema. (src/config/zod-schema.providers-core.ts:102, a1271830948b)commentaryfor Discord, Slack, and Telegram progress streaming. (src/config/schema.test.ts:305, a1271830948b)Likely related people:
src/config/zod-schema.providers-core.ts,src/config/schema.test.ts,extensions/discord/src/config-ui-hints.ts)extensions/discord/src/monitor/message-handler.draft-preview.ts,extensions/discord/src/monitor/message-handler.process.ts,.github/workflows/docker-release.yml)src/config/zod-schema.providers-core.ts,.github/workflows/docker-release.yml,scripts/release-preflight.mjs)Codex review notes: model internal, reasoning high; reviewed against a1271830948b.
ericpearson commentedon May 31, 2026
Additional local remediation attempt and deployment check:
I ran
openclaw doctor --fixinside the affected container before retrying the schema path.Doctor did make unrelated config/state repairs:
It then exited nonzero while trying to install shell completion:
After doctor, the schema mismatch still reproduced:
Output:
And the validated config write still failed:
Output:
Since doctor did not fix it, I triggered the Docker updater webhook flow from this deployment. The update worker run id is:
The worker is configured to redeploy via the existing updater webhook, then perform a delayed post-restart check of:
If the fresh pull/redeploy still reproduces the same failure, that should prove this is reproducible from the current GHCR Docker artifact rather than local config drift. The next useful validation from maintainers would be to pull the same GHCR image into a clean container, configure Discord, set
channels.discord.streaming.mode=progressplusstreaming.progress.commentary=true, and verify with an actual Discord turn that the progress draft shows commentary before final delivery.ericpearson commentedon May 31, 2026
@clawsweeper re-review
Please review this fresh packaging-specific report. This is intentionally scoped to the running GHCR/Docker artifact, not to the already-merged source implementation in #85200.
Since filing, I also ran
openclaw doctor --fix; it made unrelated repairs but the same schema failure still reproduces:and:
I then triggered the deployment updater webhook so the current deployment pulls/redeploys. If the post-redeploy check still shows
hasCommentary=false, that should be strong evidence this is reproducible from the current GHCR artifact rather than local config drift.The requested maintainer validation is: pull the relevant GHCR image, run
node /app/openclaw.mjs config schema, confirm whetherchannels.discord.streaming.progress.properties.commentaryexists, then try a Discord config/run withstreaming.mode=progressandstreaming.progress.commentary=true.clawsweeper commentedon May 31, 2026
🦞🧹
ClawSweeper re-review requested.
I asked ClawSweeper to review this item again.
Action: item re-review queued (workflow
sweep.yml, eventrepository_dispatch).Result: the existing ClawSweeper review comment will be edited in place when the review finishes.
Re-review progress:
ericpearson commentedon May 31, 2026
Post-redeploy verification from the affected deployment:
The Docker updater webhook completed successfully from this deployment:
I then rechecked the live container after the delayed post-restart window.
Timestamp:
Running version and package:
The emitted runtime config schema still does not include the Discord progress commentary key:
Output:
The strict config write still fails after redeploy:
Output:
So the remediation chain is now:
openclaw doctor --fixwas run and made unrelated repairs, but did not fix the schema.OpenClaw 2026.5.28.channels.discord.streaming.progress.commentary.This continues to look like the currently deployed GHCR image/package artifact is stale or mixed for the generated runtime schema, even though source/tag contain the feature.
ericpearson commentedon May 31, 2026
@clawsweeper re-review
Please re-review with the new post-redeploy evidence now attached in #88788 (comment).
The affected deployment has now gone through the full local remediation chain:
openclaw doctor --fixran and made unrelated repairs.This should move the issue from "possible local config drift" to "reproducible after a successful redeploy from the current deployed GHCR/container path." Please re-check whether maintainers need to pull/inspect the published GHCR image/manifest, not just source or the v2026.5.28 tag.
clawsweeper commentedon May 31, 2026
🦞👀
ClawSweeper picked this up.
Command router queued. I will update this comment with the next step.
Re-review progress:
ericpearson commentedon May 31, 2026
@clawsweeper re-review
One important clarification about the deployment/update path, since this issue is specifically about the published container artifact:
This deployment is not rebuilding OpenClaw from local source and it is not using an in-container git checkout. The container reports the normal Docker behavior for
gateway update.runasnot-git-install, so updates are handled by an external updater webhook:That webhook runs outside the OpenClaw app container and redeploys the stack. In this run it completed with:
After that external redeploy/pull path, the restarted app container still reported:
and still rejected:
So the SHA/version/runtime evidence here should be treated as coming from the deployed GHCR Docker artifact selected by the updater, not from a local source checkout. The missing remaining evidence is the exact GHCR image digest/manifest from the host or registry side; the affected app container does not have Docker socket access, so it can verify runtime contents but cannot inspect its own pulled image digest.
Please re-review with that in mind: the requested maintainer repro is to pull/inspect the published GHCR image/manifest selected for
latest/2026.5.28, then run the container's ownnode /app/openclaw.mjs config schemasmoke and confirm whether the emitted schema includeschannels.discord.streaming.progress.properties.commentary.clawsweeper commentedon May 31, 2026
🦞👀
ClawSweeper picked this up.
Command router queued. I will update this comment with the next step.
Re-review progress:
14 remaining items