| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| caprock_0.44.3_darwin_amd64.tar.gz | 2026-09-02 | 8.7 MB | |
| caprock_0.44.3_windows_arm64.zip | 2026-09-02 | 8.1 MB | |
| checksums.txt | 2026-09-02 | 600 Bytes | |
| caprock_0.44.3_darwin_arm64.tar.gz | 2026-09-02 | 8.3 MB | |
| caprock_0.44.3_linux_amd64.tar.gz | 2026-09-02 | 8.6 MB | |
| caprock_0.44.3_linux_arm64.tar.gz | 2026-09-02 | 8.0 MB | |
| caprock_0.44.3_windows_amd64.zip | 2026-09-02 | 8.8 MB | |
| README.md | 2026-09-02 | 1.0 kB | |
| v0.44.3 source code.tar.gz | 2026-09-02 | 5.7 MB | |
| v0.44.3 source code.zip | 2026-09-02 | 6.0 MB | |
| Totals: 10 Items | 62.1 MB | 0 | |
Fixed
- A session you were still working in could close itself. 0.44.0 started
listening to Claude Code's
SessionEndhook so a session ends the moment it ends rather than whenever the staleness sweep next runs. ButSessionEnddoes not only mean "the user left": it also fires on/clear, on Escape at the prompt, and with an unhelpfulother. All of them were being treated as the end. One session on the machine this was found on had been running for six days when a/clearretired it in the dashboard.
Only a reason that actually means an exit ends the session now. A /clear is
recorded as the context reset it is, so it still appears in the timeline. A
process that really exits is unaffected — that path records the exit code and
never consulted the reason.
The two mistakes here are not symmetric, and the fix leans accordingly: a session wrongly left open closes itself within the hour, while one wrongly closed vanishes while its owner is still typing into it.