PR #63583: feat(hooks): respect userTimezone in session-memory timestamps #63454

• Repository: openclaw/openclaw
• Author: jepson-liu
• State: closed | merged: False
• Link: https://github.com/openclaw/openclaw/pull/63583

Description (problem / solution / changelog)

Summary

• Problem: The bundled session-memory hook always derived the dated filename prefix and # Session header from UTC (toISOString()), so users with agents.defaults.userTimezone set still saw UTC in saved memory files—misaligned with the system prompt and Issue #63454.
• Why it matters: Operators in non-UTC zones could not line up memory filenames/headers with local conversation time.
• What changed: session-memory now uses resolveUserTimezone plus shared formatDateStampInTimezone / formatWallClockHmsInTimezone for the date prefix, header line, and fallback HHMM slug; header ends with the resolved IANA zone id. formatDateStampInTimezone is exported via the memory host runtime / openclaw/plugin-sdk/memory-core and reused by extensions/memory-core/src/flush-plan.ts to avoid duplicate date logic.
• What did NOT change (scope boundary): Pre-compaction memory flush behavior beyond sharing the same date-stamp helper; no new config keys; no gateway protocol or channel changes.

Change Type (select all)

• Bug fix
• Feature
• Refactor required for the fix
• Docs
• Security hardening
• Chore/infra

Scope (select all touched areas)

• Gateway / orchestration
• Skills / tool execution
• Auth / tokens
• Memory / storage
• Integrations
• API / contracts
• UI / DX
• CI/CD / infra

Linked Issue/PR

• Fixes #63454

User-visible / Behavior Changes

• session-memory hook: saved files under memory/YYYY-MM-DD-<slug>.md use the calendar day in agents.defaults.userTimezone (fallback: host zone or UTC). First line is # Session: <date> <time> <IANA zone> instead of ... UTC.
• Plugin SDK: formatDateStampInTimezone is now part of the openclaw/plugin-sdk/memory-core export surface (additive).

Security Impact (required)

• New permissions/capabilities? No
• Secrets/tokens handling changed? No
• New/changed network calls? No
• Command/tool execution surface changed? No
• Data access scope changed? No

Repro + Verification

Environment

• OS: Linux (CI / local dev)
• Runtime/container: Node 22+ / Vitest
• Relevant config (redacted): agents.defaults.userTimezone: "Asia/Jakarta" (or any valid IANA zone)

Steps

1. Enable session-memory and set agents.defaults.userTimezone to a non-UTC zone (e.g. Asia/Jakarta).
2. Trigger /new or /reset so the hook writes a memory file.
3. Open the new file under memory/.

Expected

• Filename date and # Session header reflect local wall time in the configured timezone; header suffix matches the IANA zone name.

Actual (before)

• UTC-based date/time and UTC suffix in the header.

Evidence

• Failing test/log before + passing after
• Trace/log snippets
• Screenshot/recording
• Perf numbers (if relevant)

(Add in PR if useful: new Vitest case in src/hooks/bundled/session-memory/handler.test.ts pins 2026-04-09T00:44:38.000Z with Asia/Jakarta and asserts 2026-04-09 + 07:44:38 + Asia/Jakarta.)

Human Verification (required)

• Verified scenarios: pnpm test src/hooks/bundled/session-memory/handler.test.ts; pnpm test extensions/memory-core/index.test.ts -t buildMemoryFlushPlan; pnpm check via scripts/committer on commit.
• Edge cases checked: invalid/missing timezone falls back via existing resolveUserTimezone behavior.
• What I did not verify: Full `pnpm

Changed files

• CHANGELOG.md (modified, +1/-0)
• docs/automation/hooks.md (modified, +1/-1)
• extensions/memory-core/src/flush-plan.ts (modified, +1/-16)
• src/agents/date-time.ts (modified, +44/-0)
• src/hooks/bundled/session-memory/HOOK.md (modified, +2/-2)
• src/hooks/bundled/session-memory/handler.test.ts (modified, +25/-0)
• src/hooks/bundled/session-memory/handler.ts (modified, +12/-8)
• src/memory-host-sdk/runtime-core.ts (modified, +1/-0)
• src/plugin-sdk/memory-core.ts (modified, +1/-0)