chore: sync upstream through c3a36af872

[andrewcai8]

Brings in the 23 upstream commits since #207. Merge with Create a merge commit, never squash, so the next sync starts from this merge.

Upstream changes that matter for andrew.megpt.app:

• fix(server): a dropped HTTP/2 connection no longer crashes the server pingdotgg/t3code#18378: a dropped HTTP/2 connection no longer crashes the server.
• fix(server): fail runs whose provider stream ends without a terminal pingdotgg/t3code#19131: a run whose provider stream ends without a terminal event now fails instead of hanging.
• fix(acp): one malformed session update no longer ends the session pingdotgg/t3code#20015: one malformed ACP session update no longer ends the session.

The rest is web and mobile polish, Windows install fixes, Beta branding, and GitHub GraphQL moving onto effect/graphql (pingdotgg#19224).

Conflicts:

• apps/server/package.json: took upstream's catalog undici 8.11.2, replacing the fork's 7.29.1 pin.
  • The fork pinned 7 (fix(cloud): deleting an E2B box stops its sandbox again #163) so the bundled E2B SDK would not resolve Cursor's undici 5.
  • The SDK also loads undici 8, through its optional undici8 alias, and prefers it on Node 22.19 and newer.
  • runtimeDependencies.test.ts now accepts either major the SDK loads, and still fails on undici 5.
  • Checked against E2B: an h2 agent on undici 8.11.2 killed a throwaway sandbox with a body-less DELETE in 179 ms. That is the request fix(cloud): deleting an E2B box stops its sandbox again #163 saw hang for 10 s.
• ci.yml: kept both steps: the fork's case-collision check and upstream's GraphQL generation check.
• docs/operations/release.md: kept the fork's publication section and upstream's Beta note.
• scripts/lib/cli-external-packages.test.ts: kept the fork's @bufbuild/* and @connectrpc/* externals for the Namespace client.
• pnpm-lock.yaml: regenerated with vp i.

Ran locally: outboundHttp.test.ts (upstream's version, unchanged), runtimeDependencies.test.ts and cli-external-packages.test.ts.

🤖 Generated with Claude Code

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	4.9 KiB	5.0 KiB	+23 B (+0.5%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.2 KiB	1.2 KiB	+23 B (+1.9%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	20.8 KiB	20.9 KiB	+41 B (+0.2%)	29.3 KiB	✅
Codex	Live turn messages	1	2	+1 (+100.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	−24 B (−0.5%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.2 KiB	1.2 KiB	−24 B (−2.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.2 KiB	21.2 KiB	−41 B (−0.2%)	29.3 KiB	✅
Claude	Live turn messages	2	1	−1 (−50.0%)	8	✅

Baseline: 759cd75 · PR result: ebb4919 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.