fix(gateway): handoffs fail closed and load secrets off-loop; Discord slash commands honor profile_routes (#97693 #100014 #69178 #91633, salvage #97694 #100049)

[teknium1]

Summary

Three multiplex-gateway bugs in one branch: a secondary-profile handoff whose config failed to load silently delivered through the primary's home channel; the handoff watcher (and per-message handlers) entered the profile secret scope synchronously on the event loop, stalling every adapter during slow secret reads; and native Discord slash commands built their SessionSource without guild_id/parent_chat_id, so guild/channel profile_routes never matched and /new, /reset, /model, /status ran against the default profile.

Changes

• _process_handoff: config-load failure → error + raise (row marked failed) instead of falling back to self.config (fix(gateway): fail closed when a secondary profile handoff cannot load its config #97694, @Adolanium, cherry-picked).
• _load_profile_secret_scope + _async_profile_runtime_scope: secret scope loaded via asyncio.to_thread, then the existing sync scope is entered with prepared_secret_scope=; watcher reclaim/tick use it (fix(gateway): keep multiplex handoffs responsive during slow secret reads #100049, @fangliquanflq, cherry-picked).
• Class widening: the three async per-message scope entries (_make_profile_message_handler, _make_default_profile_message_handler, _prepare_inbound_message_text_scoped) now use the async scope too.
• Discord _build_slash_event / _dispatch_thread_session pass guild_id (new _interaction_guild_id, mirrors on_message) and parent_chat_id to build_source.
• Tests: 1 fail-closed handoff test, 1 loop-responsiveness test, 1 slash-routing parity test (+ existing watcher tests adapted to the async scope).

Validation

Scenario	origin/main	this branch
Secondary handoff, profile config unloadable	create_handoff_thread on PRIMARY home '111', row completed	RuntimeError, no adapter call, row failed
Watcher tick with 600ms secret hydration	loop stalled 601 ms	21 ms
Discord /new in guild-routed channel	profile=None, reset key agent:main:…	profile='work', key agent:work:… (idle + busy)
/status in thread	guild_id/parent_chat_id None → no route	guild_id='1', parent_chat_id='100' → routed
Sabotage: each regression test fails with its fix reverted. Targeted suites: 70 passed; 34-file seam sweep: 586 passed.

Credits

@Adolanium (#97694 merged as-is), @fangliquanflq (#100049, first submitter), @Tranquil-Flow (#100076, same mechanism, co-author), @Sora-bluesky (#69242, first fix for #69178, co-author), @jondgilbert (#91634 / #91633, co-author), @tensorbit89-netizen (#31102, first report), @vKongv (#85355 diagnosis; its residual /new symptom is the guild_id gap fixed here).

Probes kept at /tmp/mux_sweep/hds/ (probe_failclosed.py, probe_watcher.py, probe_discord_slash.py, probe_reset_key.py, probe_85355.py). Baseline worktree /tmp/salv-hds-base (detached origin/main) also left in place.

Fixes #97693
Fixes #100014
Fixes #69178
Fixes #91633

## Infographic

![mux-handoff-discord-slash](https://v3b.fal.media/files/b/0aa8cde9/NJuZa3RsvwDHXZIDrBnou_BOEjghDB.png)

[github-actions]

૮ >ﻌ< ა ci review

_{ran on 8c1ca9b — fix(discord): native slash commands honor guild/channel prof}

⚠️ Warnings

OSV vulnerability scan · View job

13 known vulnerabilities found in pinned dependencies.

• CVE-2026-67213 in website/package-lock.json
• CVE-2026-71554 in uv.lock
• CVE-2026-73088 in package-lock.json
• CVE-2026-73088 in website/package-lock.json
• GHSA-8423-8fgw-73vq in uv.lock
• CVE-2026-70608 in package-lock.json
• CVE-2026-73089 in package-lock.json
• CVE-2026-73089 in website/package-lock.json
• CVE-2026-84371 in package-lock.json
• CVE-2026-56876 in package-lock.json
• CVE-2026-70606 in package-lock.json
• GHSA-wwv5-g3v4-889x in uv.lock
• CVE-2026-71554 in uv.lock

How to fix:

Review the findings in the Security tab. Update the affected dependencies if a patched version is available.

---

debug info

CI timings

CI timings · View report · View job

Wall time 6m10s vs 5m15s (+17.5%). 8 job(s) slower, 3 faster, 3 unchanged.

• Detect affected areas: -47.0s
• Python tests / Run tests: +22.0s
• OS-specific tests / Windows-only tests: +13.0s
• OS-specific tests / macOS-only tests: +12.0s
• Python lints / Windows footguns (blocking): +8.0s