Merge branch 'dev' into feature/ssh-support

Brings dev's retry-with-backoff (the transient `provider-retry` AgentEvent the
web frontend consumes) + the LSP-dead-server per-edit-hang fix into the SSH
feature branch, alongside the SSH waves 0-5c.

All code files auto-merged cleanly (run-turn.ts, orchestrator.ts, runtime.ts,
wire/index.ts, tool-edit-file/extension.ts, run-turn.test.ts — both computerId
threading and retry-with-backoff coexist). Only tasks.md conflicted (status
section — orchestrator-resolved; both feature sections kept).

Verified post-merge: tsc -b EXIT 0, biome clean (391 files), 1730 vitest pass
+6 sshd-integration skipped (was 1690; +40 from dev's retry/LSP tests).
Wire dist rebuilt so the FE can re-sync the pinned @dispatch/wire dep and pick
up BOTH provider-retry AND the SSH Computer/defaultComputerId types.
No merge or push (into dev or otherwise).

1 files changed, 13 insertions, 0 deletions

diff --git a/packages/lsp/src/manager.ts b/packages/lsp/src/manager.ts
index 7153956..bc84479 100644
--- a/packages/lsp/src/manager.ts
+++ b/packages/lsp/src/manager.ts
@@ -134,6 +134,19 @@ export class LspManager {
 			if (existing) {
 				const state = existing.client.getState();
 				const stateError = existing.client.getStateError();
+				// A client that died or corrupted AFTER connecting flipped its
+				// own state to "error" (client.ts handleExit/markBroken). Spawn
+				// succeeded so there's no broken entry yet — seed one so the
+				// bounded-backoff path above re-spawns it, instead of reporting
+				// error forever (and so getDiagnostics' "connected" filter skips
+				// it, avoiding a per-edit hang on the corpse).
+				if (state === "error" && !this.broken.has(key)) {
+					this.broken.set(key, {
+						configFingerprint: configFingerprint(server),
+						brokenAt: this.now(),
+						error: enrichError(server, stateError ?? "server unavailable"),
+					});
+				}
 				const status: LspServerStatus = {
 					id: server.id,
 					name: server.name,