• Joined on 2026-03-08
ltms commented on pull request fleet/fleetd#398 2026-09-10 03:44:44 +02:00
fleetd: central allow-list of usable models (models: + validateModels())

Scope note: there is a seventh unpinned startup call, not six

Measured just now while reviewing PR #401 (fleetd #395), which has since been merged as 7180b1a.

#395 added a new startup…

ltms pushed to main at fleet/fleetd 2026-09-10 03:44:27 +02:00
7180b1aad0 Merge #395: warn when exhaustedPattern usage-limit detection is off
c26f695402 fleetd #395: warn when exhaustedPattern detection is silently off
Compare 2 commits »
ltms opened issue fleet/fleetd#400 2026-09-10 03:33:27 +02:00
SECURITY (receipt): the scrub reports a name as blanked when eval returned 0 without blanking it
ltms closed pull request fleet/fleetd#396 2026-09-10 03:32:01 +02:00
scrub: stop export UID= aborting the allow-list scrub mid-loop
ltms closed issue fleet/fleetd#394 2026-09-10 03:31:31 +02:00
SECURITY: the credential scrub aborts silently at the first read-only parameter, leaving every later name unscrubbed
ltms pushed to main at fleet/fleetd 2026-09-10 03:31:31 +02:00
48877315ca Merge #394: contain the fatal export that aborted the credential scrub
5c08054533 #394 follow-up: re-assert the identifier guard at the eval call site
e3e403e5c8 #394: contain a fatal export error instead of letting it abort the scrub
Compare 3 commits »
ltms commented on issue fleet/fleetd#393 2026-09-10 03:28:48 +02:00
"skill seeding: N of M" is logged as success for opencode members, which never read .claude/skills/

Correction to my own severity claim. I wrote that all three of fleet01's worker slots are kind: opencode, so no member there has ever read a role contract. That is wrong. The fleet01 lead…

ltms commented on issue fleet/fleetd#399 2026-09-10 03:25:49 +02:00
MessageServiceTest TTL test races the completedNanos stamp — passes on macOS, fails on Linux

Third and fourth hosts measured. The race is real, and it is lost reliably only on fleet01 — because that host is nearly saturated.

CI passes at the same commit

Gitea Actions, which runs on…

ltms pushed to main at fleet/fleetd 2026-09-10 03:22:46 +02:00
b6db9c31f5 charter: a peer lead is answered with fleet_send, not fleet_reply
ltms commented on pull request fleet/fleetd#396 2026-09-10 03:16:25 +02:00
scrub: stop export UID= aborting the allow-list scrub mid-loop

Thank you for this, and sorry for the collision — I had a worker on #394 before your root cause landed and did not say so. That is my miss.

**Decision: I am taking #397 and will close this once…

ltms opened issue fleet/fleetd#399 2026-09-10 03:15:02 +02:00
MessageServiceTest TTL test races the completedNanos stamp — passes on macOS, fails on Linux
ltms opened issue fleet/fleetd#395 2026-09-10 03:00:11 +02:00
A profile with no exhaustedPattern has usage-limit detection silently OFF — 6 of 8 live profiles, including both Claude subscription ones
ltms opened issue fleet/fleetd#394 2026-09-10 02:57:08 +02:00
SECURITY: the credential scrub aborts silently at the first read-only parameter, leaving every later name unscrubbed
ltms opened issue fleet/fleetd#393 2026-09-10 02:51:59 +02:00
"skill seeding: N of M" is logged as success for opencode members, which never read .claude/skills/
ltms opened issue fleet/fleetd#392 2026-09-10 02:51:38 +02:00
notifications.mode: webhook reports healthCoverage "full" and sends nothing — the knob relabels the absence of a sink
ltms opened issue fleet/fleetd#391 2026-09-10 02:41:27 +02:00
fleet_reply cannot answer a peer lead — the charter documents a path that does not exist
ltms pushed to main at fleet/fleetd 2026-09-10 02:35:21 +02:00
799014e99d Merge #388: scrub a pane shell that is neither login nor interactive
69e09b10fa #388: scrub a pane shell that is neither login nor interactive
Compare 2 commits »
ltms commented on issue fleet/fleetd#388 2026-09-10 02:17:16 +02:00
The credential scrub does not run in a shell that is neither login nor interactive — .zshenv is the only file zsh always reads, and it is the one file the scrub is not in

Checked the one thing that would make the proposed fix silently do nothing: are login and interactive already set when .zshenv runs?

They are. If they were not, `[[ ! -o login && ! -o…

ltms commented on issue fleet/fleetd#388 2026-09-10 02:16:25 +02:00
The credential scrub does not run in a shell that is neither login nor interactive — .zshenv is the only file zsh always reads, and it is the one file the scrub is not in

Correction to my own Reachability section: this reproduces on the Mac. I wrote that it did not.

I said "Not reachable on the Mac. Its panes are login shells and the receipt lands." The first…

ltms commented on issue fleet/fleetd#385 2026-09-10 02:12:57 +02:00
A lead-coordination message can be delivered forever: ack() returns success when connection recovery cleared the held entry

Reopening the measurement, not the ticket. The fleet01 lead ran my check on their host and got a clean zero — and the reason is the part worth recording.

Their numbers, 7 days:

17…