ltms - Gitea

ltms closed issue fleet/fleetd#201

2026-09-03 06:58:04 +02:00

Backend-error classification needs a real mechanism, not hard-coded strings (#164 points 3 and 4)

ltms commented on issue fleet/fleetd#201

2026-09-03 06:57:58 +02:00

Backend-error classification needs a real mechanism, not hard-coded strings (#164 points 3 and 4)

Done and merged. All five units are on main.

ltms closed pull request fleet/fleetd#243

2026-09-03 06:57:47 +02:00

fleetd #241: suppress echoed fallback briefs

ltms commented on issue fleet/fleetd#241

2026-09-03 06:57:44 +02:00

Completion fallback can hand the lead back its own brief as the member's "report"

Correction to my comment above: the PR for this work is #243, not #245. #245 is the #134 tool-surface PR. The merge commit 5cf3ca9 carries the same wrong number in its message; it is pushed,…

ltms closed issue fleet/fleetd#241

2026-09-03 06:57:19 +02:00

Completion fallback can hand the lead back its own brief as the member's "report"

ltms commented on issue fleet/fleetd#241

2026-09-03 06:57:14 +02:00

Completion fallback can hand the lead back its own brief as the member's "report"

Fixed and merged to main as 5cf3ca9 (PR #245).

CompletionResolver.echoesInjectedBrief now refuses a scrape that is just the injected brief read back, so a silent member no longer looks like…

ltms closed pull request fleet/fleetd#246

2026-09-03 06:57:05 +02:00

fleetd #201/#227 unit 5: wire errorPattern, cool-off spawn gate, and fleet views

ltms opened issue fleet/fleetd#248

2026-09-03 06:55:33 +02:00

Fleetd's composition root is untested: a feature can be silently unwired and every test stays green

ltms pushed to main at fleet/fleetd

2026-09-03 06:54:19 +02:00

5cf3ca9a89 fleetd #241: never hand the lead back its own brief as the member's report

ac474981e4 fleetd #201/#227 Unit 5: wire the backend-error cool-off into config, placement and the MCP surface

ba04b2359b fleetd #201/#227 unit 5: wire errorPattern, cool-off spawn gate, and fleet views

3437d6313d fleetd #241: bound the echo match so a real report is never swallowed

321d8dcbb5 fleetd #241: suppress echoed fallback briefs

Compare 5 commits »

ltms commented on issue fleet/fleetd#148

2026-09-03 06:50:44 +02:00

CB-634: the parity overlay copies .env/.envrc into worker worktrees — the same shape as CB-633

Decision on points 1 and 2 (lead), with the evidence the ticket asked for first

Live exposure: none today. I checked this host on 2026-09-03:

ls -la .env .envrc in the primary…

ltms commented on issue fleet/fleetd#149

2026-09-03 06:48:16 +02:00

A claude-code member spawned into a fresh worktree blocks forever on the workspace trust dialog

Fixed and merged to main as 2e5b63f (PR #244).

ClaudeCodeLauncher now seeds projects.<cwd>.hasTrustDialogAccepted and hasCompletedProjectOnboarding in ~/.claude.json (or `configDir/.cl…

ltms opened issue fleet/fleetd#247

2026-09-03 06:47:57 +02:00

Seeding ~/.claude.json can still lose an external writer's change

ltms closed pull request fleet/fleetd#244

2026-09-03 06:47:40 +02:00

fleetd #149: seed the workspace-trust dialog before a claude-code spawn

ltms commented on pull request fleet/fleetd#244

2026-09-03 06:47:34 +02:00

fleetd #149: seed the workspace-trust dialog before a claude-code spawn

Merged into main as 2e5b63f (round 2 head 743377d). Closing this PR — the branch is in main, not dropped.

How I checked it (lead, in an isolated worktree):

ltms pushed to main at fleet/fleetd

2026-09-03 06:45:11 +02:00

2e5b63f6f6 fleetd #149: seed the workspace-trust entry before a claude-code spawn

743377d6cd fleetd #149 review round 2: make the trust-dialog seed atomic and lock-protected

a89dcc9b7e fleetd #149: seed the workspace-trust entry before a claude-code spawn

Compare 3 commits »

ltms commented on pull request fleet/fleetd#243

2026-09-03 06:39:39 +02:00

fleetd #241: suppress echoed fallback briefs

Round 2 pushed as 3437d63. Verified, but not yet merged — it must land after #201/#227 Unit 5, see the collision note below.

The round-1 defect and the fix

Round 1 used plain…

ltms pushed to worker/cb241-fallback-echo-1175e9-11 at fleet/fleetd

2026-09-03 06:39:15 +02:00

3437d6313d fleetd #241: bound the echo match so a real report is never swallowed

ltms closed pull request fleet/fleetd#245

2026-09-03 06:32:41 +02:00

fleetd #134: make tool-surface neutralization visible to the daemon and worker

ltms closed issue fleet/fleetd#134

2026-09-03 06:32:34 +02:00

CB-628: a tracked file neutralized by the worktree overlay can never be edited by a worker, and nothing says so

ltms commented on issue fleet/fleetd#134

2026-09-03 06:32:28 +02:00

CB-628: a tracked file neutralized by the worktree overlay can never be edited by a worker, and nothing says so

All three points are done. Merged to main in 205ad82 (points 1 and 2, PR #245) and 5952d55 (point 3, applied by me).

Point 1 — the daemon says it

tool-surface isolation: neutralized…