This release fixes issues for SREs watching stability and regressions.
✓ No known CVEs patched in this version
Topics
+14 more
ReleasePort's take
Light signalVersion v0.4.5 hardens writable file locking to prevent concurrent writers from corrupting `.rememb/entries.json`.
Why it matters: Patch to v0.4.5 immediately if your workload writes concurrently to `.rememb/entries.json` to avoid data corruption.
Summary
AI summaryHardened writable file locking to prevent concurrent writers from corrupting .rememb/entries.json.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Bugfix | Medium |
Treat mutating file modes as exclusive locks in the storage helper Treat mutating file modes as exclusive locks in the storage helper Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Other | Medium |
Cover r+ write-lock exclusivity with a multiprocessing regression test Cover r+ write-lock exclusivity with a multiprocessing regression test Source: llm_adapter@2026-05-21 Confidence: low |
— |
Full changelog
Highlights
- Prevented concurrent writers from corrupting
.rememb/entries.jsonduring large delete and consolidation operations. - Hardened writable file locking so every mutating mode, including
r+, is treated as an exclusive write lock. - Added a regression test that proves a second process cannot acquire a writable lock while another write operation is still in progress.
Included changes
- fix: treat mutating file modes as exclusive locks in the storage helper
- test: cover
r+write-lock exclusivity with a multiprocessing regression test - chore: bump package and release metadata to 0.4.5
Validation
- Ran
pytest tests/test_store.py -q - Ran
python -m py_compile src/rememb/*.py tests/test_store.py - Built wheel and sdist with
python -m build - Verified package metadata with
python -m twine check dist/*
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About LuizEduPP/rememb
Persistent memory for AI agents with sectioned entries (project/user/context/etc), semantic search, CLI, and per-project scope. Local JSON, zero config, no server required.
Related context
Related tools
Beta — feedback welcome: [email protected]