ltms
0 Followers · 0 Following
  • Joined on 2026-03-08
Repositories Projects Packages Public Activity Starred Repositories
ltms commented on issue fleet/fleetd#404 2026-09-10 04:21:30 +02:00
exhaustionDetectionArmed reads the LIVE config while detection reads the STARTUP snapshot — after a reload it reports armed when detection is still off

Merged as PR #406, with one test I added during review.

The fix

The armed lambda now reads the startup pattern map — exhaustedPatternsByProfile, the same map CompletionResolver…

ltms pushed to main at fleet/fleetd 2026-09-10 04:20:56 +02:00
a502ba53e0 Merge #404: the armed field reads the startup pattern map, both directions pinned
446cc11d1d fleetd #404: pin the armed field's TRUE direction too
b8aedeafcb fleetd #404: test armed startup map behavior
44c735f6f5 fleetd #404: report armed detection from startup map
Compare 4 commits »
ltms merged pull request fleet/fleetd#406 2026-09-10 04:20:55 +02:00
fleetd #404: report armed detection from startup map
ltms commented on issue fleet/fleetd#391 2026-09-10 04:18:59 +02:00
fleet_reply cannot answer a peer lead — the charter documents a path that does not exist

Merged as PR #402, in two rounds. The second round is the one worth reading.

Round 1 — the refusal itself

fleet_reply now refuses a PRIMARY caller and names the two routes that do…

ltms merged pull request fleet/fleetd#402 2026-09-10 04:18:37 +02:00
fleetd #391: refuse lead fleet replies
ltms pushed to main at fleet/fleetd 2026-09-10 04:18:37 +02:00
ddd81fe174 Merge #391: fleet_reply refuses a lead, and the role argument is now mandatory
cf8da1d5fa fleetd #391: require reply caller role
3982ace544 fleetd #391: refuse lead fleet replies
Compare 3 commits »
ltms commented on issue fleet/fleetd#399 2026-09-10 04:15:04 +02:00
MessageServiceTest TTL test races the completedNanos stamp — passes on macOS, fails on Linux

Merged as PR #405.

What I checked myself

  1. The production diff is purely additive. One new field, one hook line, one package-private test seam. I grepped the diff for expired and…
ltms merged pull request fleet/fleetd#405 2026-09-10 04:14:30 +02:00
fleetd #399: fix TTL test race — wait on completion stamp, not Phase.DONE
ltms pushed to main at fleet/fleetd 2026-09-10 04:14:30 +02:00
ae74cc081f Merge #399: wait on a real completion stamp, not on DONE
0f51d53098 fleetd #399: fix TTL test race by waiting on a real completion stamp, not DONE
Compare 2 commits »
ltms pushed to main at fleet/fleetd 2026-09-10 04:06:44 +02:00
3d61af6f6f Merge #400: the scrub receipt now measures the blank, not the attempt
c670792ffe #400: classify the blanking loop's result on the observed value, not eval's exit status
Compare 2 commits »
ltms pushed to main at fleet/fleetd 2026-09-10 04:04:09 +02:00
ed99c209ac Merge #398: a central allow-list of usable models, with the startup validators pinned
af4c88d54b t398: correct two false statements in FleetConfigValidateAllTest
b540a1744b fleetd #398 follow-up: pin the startup validators with a reflective validateAll()
e7b33fe3a0 fleetd: central allow-list of usable models (models: + validateModels())
Compare 4 commits »
ltms opened issue fleet/fleetd#408 2026-09-10 04:03:53 +02:00
Five places trust a command's exit code as proof of its effect and never read the value back — the #400 shape outside the scrub
ltms commented on pull request fleet/fleetd#403 2026-09-10 04:03:06 +02:00
#400: classify the blanking loop on the observed value, not eval exit status

Lead review — good work, and one part of it is better than what I asked for

I read the diff rather than the report. The change is minimal and right:

-                  if eval "export…
ltms opened issue fleet/fleetd#407 2026-09-10 04:01:31 +02:00
Five log-only startup reporters in Fleetd.main have unpinned call sites — deleting any one ships a green build
ltms commented on pull request fleet/fleetd#398 2026-09-10 04:00:49 +02:00
fleetd: central allow-list of usable models (models: + validateModels())

Lead review — the follow-up work was recovered from a dead worker, and it is good

First, the recovery. The worker's agent died mid-turn (herdr error [agent_not_found]) with all of this…

ltms opened issue fleet/fleetd#404 2026-09-10 03:50:18 +02:00
exhaustionDetectionArmed reads the LIVE config while detection reads the STARTUP snapshot — after a reload it reports armed when detection is still off
ltms closed issue fleet/fleetd#395 2026-09-10 03:46:47 +02:00
A profile with no exhaustedPattern has usage-limit detection silently OFF — 6 of 8 live profiles, including both Claude subscription ones
ltms commented on issue fleet/fleetd#395 2026-09-10 03:46:41 +02:00
A profile with no exhaustedPattern has usage-limit detection silently OFF — 6 of 8 live profiles, including both Claude subscription ones

Fixed and merged as 7180b1a

PR #401, merged onto main with a --no-ff merge. Merged build measured by me: Tests run: 1474, Failures: 0, Errors: 0, Skipped: 0, BUILD SUCCESS, 0…

ltms commented on pull request fleet/fleetd#401 2026-09-10 03:45:51 +02:00
fleetd #395: warn when usage-limit detection is silently off

Merged by the lead as 7180b1a (a --no-ff merge onto main, so the forge does not mark this PR merged). Closing.

What I verified myself, not from the report:

  • Merged build on main:…
ltms closed pull request fleet/fleetd#401 2026-09-10 03:45:32 +02:00
fleetd #395: warn when usage-limit detection is silently off