fix(memory): skip empty dreaming placeholders by hyspacex · Pull Request #80916 · openclaw/openclaw

hyspacex · 2026-05-12T06:04:58Z

Summary

stop REM preview from writing the No strong candidate truths surfaced placeholder under ### Possible Lasting Truths
defensively classify historical empty-result placeholders as contaminated so already-written daily memory candidates do not promote into MEMORY.md
add regression coverage for both REM preview output and short-term promotion apply filtering

Real behavior proof

Behavior or issue addressed: Empty REM days should not emit candidate-like placeholder markdown, and already-written placeholder snippets should not be promoted into long-term MEMORY.md.
Real environment tested: Harry's Mac mini, Darwin 25.4.0 arm64, Node v22.22.0, local OpenClaw checkout on branch fix/80858-dreaming-empty-placeholder.
Exact steps or command run after this patch:

node --import tsx - <<'EOF'
import fs from 'node:fs/promises';
import os from 'node:os';
import path from 'node:path';
import { __testing as dreamingTesting } from './extensions/memory-core/src/dreaming-phases.ts';
import { applyShortTermPromotions } from './extensions/memory-core/src/short-term-promotion.ts';

const preview = dreamingTesting.previewRemDreaming({ entries: [], limit: 5, minPatternStrength: 0 });
console.log('REM body contains placeholder:', preview.bodyLines.join('\n').includes('No strong candidate truths surfaced'));
console.log('REM body contains possible truths section:', preview.bodyLines.join('\n').includes('### Possible Lasting Truths'));
console.log('REM body:');
console.log(preview.bodyLines.join('\n'));

const workspaceDir = await fs.mkdtemp(path.join(os.tmpdir(), 'openclaw-80858-proof-'));
await fs.mkdir(path.join(workspaceDir, 'memory', '.dreams'), { recursive: true });
const applied = await applyShortTermPromotions({
  workspaceDir,
  minScore: 0,
  minRecallCount: 0,
  minUniqueQueries: 0,
  candidates: [{
    key: 'memory:memory/2026-04-08.md:13:13',
    path: 'memory/2026-04-08.md',
    startLine: 13,
    endLine: 13,
    source: 'memory',
    snippet: '- Candidate: Possible Lasting Truths: No strong candidate truths surfaced.',
    recallCount: 3,
    avgScore: 0.95,
    maxScore: 0.95,
    uniqueQueries: 2,
    firstRecalledAt: '2026-04-08T00:00:00.000Z',
    lastRecalledAt: '2026-04-09T00:00:00.000Z',
    ageDays: 0,
    score: 0.95,
    recallDays: ['2026-04-08', '2026-04-09'],
    conceptTags: ['candidate', 'truths'],
    components: { frequency: 1, relevance: 1, diversity: 1, recency: 1, consolidation: 1, conceptual: 1 },
  }],
});
console.log('placeholder promotions applied:', applied.applied);
console.log('MEMORY.md exists:', await fs.access(path.join(workspaceDir, 'MEMORY.md')).then(() => true, () => false));
EOF

Evidence after fix: terminal output from the real local checkout:

REM body contains placeholder: false
REM body contains possible truths section: false
REM body contains placeholder: false
REM body contains possible truths section: false
REM body line 1: ### Reflections
REM body line 2: - No strong patterns surfaced.
placeholder promotions applied: 0
MEMORY.md exists: false

Observed result after fix: Empty REM preview output no longer contains ### Possible Lasting Truths or No strong candidate truths surfaced; a historical placeholder-shaped promotion candidate applies 0 entries and does not create MEMORY.md.
What was not tested: A full scheduled dreaming cron run; validation exercised the real REM preview and promotion code paths directly with a temp workspace.

Validation

node scripts/test-projects.mjs extensions/memory-core/src/dreaming-phases.test.ts extensions/memory-core/src/short-term-promotion.test.ts
pnpm exec oxfmt --check --threads=1 extensions/memory-core/src/dreaming-phases.ts extensions/memory-core/src/dreaming-phases.test.ts extensions/memory-core/src/short-term-promotion.ts extensions/memory-core/src/short-term-promotion.test.ts
pnpm check:changed

Notes

This is intentionally narrow and does not attempt to replace the broader dreaming-contamination hardening discussed in #68774.

clawsweeper · 2026-05-12T06:08:05Z

Codex review: needs maintainer review before merge. Reviewed July 3, 2026, 11:49 PM ET / 03:49 UTC.

Summary
The PR changes memory-core REM preview output and short-term promotion filtering so empty dreaming placeholders are not emitted or promoted.

PR surface: Source +16, Tests +66. Total +82 across 4 files.

Reproducibility: yes. source-level: current main still emits the empty REM placeholder when no candidate truths are selected, and the promotion guard does not recognize that placeholder as contaminated. I did not run executable tests because this review had to keep the checkout read-only.

Review metrics: 1 noteworthy metric.

Live merge state: CONFLICTING / DIRTY at 211d0d1. Maintainers need a resolved exact head before CI, proof, and review can cover the actual merge result.

Stored data model
Persistent data-model change detected: unknown-data-model-change: extensions/memory-core/src/short-term-promotion.test.ts, vector/embedding metadata: extensions/memory-core/src/short-term-promotion.test.ts, vector/embedding metadata: extensions/memory-core/src/short-term-promotion.ts. Confirm migration or upgrade compatibility proof before merge.

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🦞 diamond lobster
Patch quality: 🐚 platinum hermit
Result: ready for maintainer review.

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

Rank-up moves:

Resolve the current-main conflict while preserving newer dreaming-contamination and fence-overlap hardening.
Rerun the focused memory-core tests and format check on the resolved exact head.

Risk before merge

[P1] GitHub reports the PR branch as CONFLICTING / DIRTY, so maintainers cannot treat the current head as merge-ready until conflict resolution is reviewed and revalidated.
[P1] The change is on the durable memory promotion path, so a bad merge resolution could either keep placeholder contamination or accidentally suppress legitimate durable candidates.

Maintainer options:

Resolve the memory-core conflict first (recommended)
Rebase or manually resolve the PR so current main's contamination, fence-overlap, and raw-snippet guards remain intact alongside the placeholder-specific fix, then rerun focused memory-core validation.
Use a clean replacement if resolution gets noisy
If this branch becomes harder to resolve than the change warrants, keep the linked issue open until a narrow replacement carries the same REM-output and promotion-filter fix with proof.

Next step before merge

[P2] Manual review is needed because the branch is conflicting; the remaining action is conflict resolution and exact-head validation, not a new independent fix-PR lane.

Security
Cleared: The diff is limited to memory-core filtering logic and tests, with no dependency, CI, secret, network, install, or code-execution surface added.

Review details

Best possible solution:

Resolve this PR against current main while preserving newer memory-core contamination guards, then land the two-sided fix that stops empty REM candidate output and rejects already-written placeholder snippets before promotion.

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

Yes, source-level: current main still emits the empty REM placeholder when no candidate truths are selected, and the promotion guard does not recognize that placeholder as contaminated. I did not run executable tests because this review had to keep the checkout read-only.

Is this the best way to solve the issue?

Yes after conflict resolution: the PR uses the narrow maintainable shape by stopping candidate-like empty REM output and rejecting already-written placeholder snippets through the existing promotion contamination gate.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 9d68f877ac3e.

Label changes

Label justifications:

P2: This is a normal-priority memory-core bugfix that prevents meaningless placeholder text from polluting durable workspace memory.
merge-risk: 🚨 session-state: The PR changes durable memory promotion filtering, and an unsafe conflict resolution could reintroduce MEMORY.md pollution or drop newer contamination guards.
rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🦞 diamond lobster and patch quality is 🐚 platinum hermit.
status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Sufficient (terminal): The PR body includes after-fix terminal output from a real local checkout invoking the production REM preview and short-term promotion apply paths with a temp workspace.
proof: sufficient: Contributor real behavior proof is sufficient. The PR body includes after-fix terminal output from a real local checkout invoking the production REM preview and short-term promotion apply paths with a temp workspace.

Evidence reviewed

PR surface:

Source +16, Tests +66. Total +82 across 4 files.

View PR surface stats

Area	Files	Added	Removed	Net
Source	2	24	8	+16
Tests	2	66	0	+66
Docs	0	0	0	0
Config	0	0	0	0
Generated	0	0	0	0
Other	0	0	0	0
Total	4	90	8	+82

What I checked:

Current main still emits placeholder: previewRemDreaming still always writes ### Possible Lasting Truths and falls back to - No strong candidate truths surfaced. when no candidate truths exist. (extensions/memory-core/src/dreaming-phases.ts:1645, 9d68f877ac3e)
Current promotion guard lacks exact placeholder predicate: isContaminatedDreamingSnippet rejects promotion markers, prompt echoes, raw snippets, and staged metadata shapes, but not the exact empty REM placeholder text. (extensions/memory-core/src/short-term-promotion.ts:414, 9d68f877ac3e)
Promotion apply path uses that guard before MEMORY.md writes: applyShortTermPromotions filters selected and rehydrated candidates through isContaminatedDreamingSnippet, so the proposed predicate sits on the durable-memory path. (extensions/memory-core/src/short-term-promotion.ts:2408, 9d68f877ac3e)
Docs contract supports the bug boundary: The Dreaming docs say REM never writes to MEMORY.md and dreaming-generated diary/report artifacts are excluded from short-term promotion. Public docs: docs/concepts/dreaming.md. (docs/concepts/dreaming.md:58, 9d68f877ac3e)
Latest release still contains the issue: Tag v2026.6.11 still contains the same empty REM placeholder fallback and no isEmptyDreamingCandidatePlaceholder guard in short-term promotion. (extensions/memory-core/src/dreaming-phases.ts:1642, e085fa1a3ffd)
Live PR state blocks merge readiness: GitHub reports this PR open and unmerged with mergeable: CONFLICTING and mergeStateStatus: DIRTY at head 211d0d14fb475d7fd84281b0f7576c8f7e91e5ca. (211d0d14fb47)

Likely related people:

mbelinky: Merged PR metadata and git show show feat(memory-core): add REM preview and safe promotion replay #61540 added REM preview/candidate-truth output, including the empty placeholder fallback this PR removes. (role: introduced REM preview behavior; confidence: high; commits: 79348f73c8b6; files: extensions/memory-core/src/dreaming-phases.ts, extensions/memory-core/src/short-term-promotion.ts)
solomonneas: Merged PR metadata shows fix(memory-core): prevent staged dream candidates from leaking into MEMORY.md #68774 hardened the same short-term promotion contamination helper this PR extends. (role: recent adjacent filter contributor; confidence: high; commits: 314903417fcc; files: extensions/memory-core/src/short-term-promotion.ts, extensions/memory-core/src/short-term-promotion.test.ts)
gumadeiras: Merged PR metadata shows memory: block dreaming self-ingestion #66852 added earlier dreaming self-ingestion guards in the same promotion-filtering area. (role: adjacent dreaming self-ingestion contributor; confidence: medium; commits: 0c4e0d703023; files: extensions/memory-core/src/dreaming-phases.ts, extensions/memory-core/src/short-term-promotion.ts, extensions/memory-core/src/short-term-promotion.test.ts)
SebTardif: Merged PR metadata shows fix(memory-core): filter REM dreaming candidates to light-staged entries #86302 recently changed REM staging and promotion behavior in the same memory-core files after this PR opened. (role: recent REM/promotion area contributor; confidence: medium; commits: 8b42771aabd2; files: extensions/memory-core/src/dreaming-phases.ts, extensions/memory-core/src/short-term-promotion.ts, extensions/memory-core/src/short-term-promotion.test.ts)

What the crustacean ranks mean

🦀 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.

How this review workflow works

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.

openclaw-barnacle · 2026-05-27T04:52:41Z

This pull request has been automatically marked as stale due to inactivity.
Please add updates or it will be closed.

clawsweeper · 2026-05-27T05:00:19Z

ClawSweeper PR egg: ✨ hatched 🌱 uncommon Moonlit Shellbean. Rarity: 🌱 uncommon. Trait: sparkles near resolved comments.

Details

Share on X: post this hatch
Copy: My PR egg hatched a 🌱 uncommon Moonlit Shellbean in ClawSweeper.
Hatchability:

Merged PRs are hatchable.
Open PRs are hatchable when they are status: 👀 ready for maintainer look, status: 🚀 automerge armed, or labeled clawsweeper:automerge.
Closed unmerged PRs are hatchable only when one of those hatchable labels is still present in the durable record.

About:

Eggs appear after real-behavior proof passes. They are collectible flavor only.
Review momentum changes the shell state: follow-up work warms it, re-review makes it wobble, and a clean final review lets it hatch.
The hatch is seeded from this repository and PR number, so the same PR keeps the same creature; the reviewed head SHA can only change safe visual details.
Rarity is just collectible sparkle: 🥚 common, 🌱 uncommon, 💎 rare, ✨ glimmer, and 🌈 legendary.

openclaw-barnacle · 2026-07-19T04:30:06Z

This pull request has been automatically marked as stale due to inactivity.
Please add updates or it will be closed.

clawsweeper · 2026-07-19T09:54:06Z

ClawSweeper status: review started.

I am starting a fresh review of this pull request: fix(memory): skip empty dreaming placeholders This is item 1/1 in the current shard. Shard 19/22.

This placeholder means the worker is alive and reading the current context. I will edit this same comment with the actual review when the claws are done clicking.

Crustacean status: shell secured, claws on keyboard, evidence pebbles being sorted.

fix(memory): skip empty dreaming placeholders

211d0d1

hyspacex mentioned this pull request May 12, 2026

Dreaming pipeline promotes empty-result placeholder "No strong candidate truths surfaced" into MEMORY.md #80858

Open

openclaw-barnacle Bot added extensions: memory-core Extension: memory-core triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. size: S labels May 12, 2026

openclaw-barnacle Bot added proof: supplied External PR includes structured after-fix real behavior proof. and removed triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. labels May 12, 2026

clawsweeper Bot added the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 12, 2026

This was referenced May 14, 2026

[Bug]: Dreaming plugin: stopword list too narrow and no filter for cron-triggered sessions #68449

Closed

fix(memory): skip placeholder short-term promotions #82028

Closed

openclaw-barnacle Bot added the stale Marked as stale due to inactivity label May 27, 2026

openclaw-barnacle Bot removed the stale Marked as stale due to inactivity label May 28, 2026

clawsweeper Bot added the merge-risk: 🚨 session-state 🚨 May lose, corrupt, stale, or mis-associate session, agent, or context state. label Jun 14, 2026

This was referenced Jun 19, 2026

fix(memory-core): skip markdown placeholder snippets in short-term promotion (#80582) [AI-assisted] #94985

Closed

fix(memory): skip raw snippets during promotion #94636

Merged

clawsweeper Bot mentioned this pull request Jun 30, 2026

Memory: skip markdown placeholder snippets during short-term promotion #80582

Closed

anyech mentioned this pull request Jul 15, 2026

fix(memory): stop deep promotion from growing MEMORY.md #108044

Closed

openclaw-barnacle Bot added the stale Marked as stale due to inactivity label Jul 19, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix(memory): skip empty dreaming placeholders#80916

fix(memory): skip empty dreaming placeholders#80916
hyspacex wants to merge 1 commit into
openclaw:mainfrom
hyspacex:fix/80858-dreaming-empty-placeholder

hyspacex commented May 12, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 12, 2026 •

edited

Loading

Uh oh!

openclaw-barnacle Bot commented May 27, 2026

Uh oh!

clawsweeper Bot commented May 27, 2026 •

edited

Loading

Uh oh!

openclaw-barnacle Bot commented Jul 19, 2026

Uh oh!

clawsweeper Bot commented Jul 19, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Uh oh!

Conversation

hyspacex commented May 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Real behavior proof

Validation

Notes

Uh oh!

clawsweeper Bot commented May 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openclaw-barnacle Bot commented May 27, 2026

Uh oh!

clawsweeper Bot commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openclaw-barnacle Bot commented Jul 19, 2026

Uh oh!

clawsweeper Bot commented Jul 19, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

hyspacex commented May 12, 2026 •

edited

Loading

clawsweeper Bot commented May 12, 2026 •

edited

Loading

clawsweeper Bot commented May 27, 2026 •

edited

Loading