llm: routes that name a host, a backup and a window, and a profile that reads them ...

`llm.json` replaces `OLLAMA_BACKENDS_FILE`. It has targets (host and key
environment variable) and routes, where a route is an ordered chain of
`model @ target` entries, each with its own `context_window` and options. The
`default` route now starts on `glm-5.3-flash:cloud` and backs it with
`qwen3.6:35b-128k` on the home host, both at 131072 tokens, so a failover no
longer changes the budget the compressor and `num_ctx` are accounted against.
Embeddings get their own `embedding` route instead of a pinned host.

The engine classifies a failure (network, quota, auth, server, missing model),
cools that entry and moves the call down the chain; a background probe watches
only degraded targets and lets the next turn return to the primary once it is
healthy — it never interrupts a running turn. A profile joins a route with
`llm_route`, and the loader now actually reads it: `llm_route` and
`context_window` were declared on `AgentProfile` and written into every
`config.json` while the loader skipped both, the same silent no-op that
`is_admin_only` used to be, so every profile ran the legacy model list at the
global window. The writer emits them too, or an admin edit would erase the
route from disk.

The turn's resolved entry travels to the UI: the first chunk is stamped with
model and target, `model_info` carries `fallback` when a backup answered, and
the chat header shows a chip only then. `serving_primary` compares the target
as well as the model name, because a route may run one model on two hosts.
`/health` gains an `llm` block with the active entry, target states and missing
models.

`deploy/ollama/` keeps the local model's definition honest: MTP off (slower at
every context on this host), `presence_penalty 0` (the base model's 1.5 breaks
agentic tool calls), and `num_gpu 40` — Ollama's automatic fit claims a 128k
f16 slot fits and then dies mid-prefill with CUDA OOM; 40 of 42 layers leaves
~400 MiB and decodes at ~77 tok/s, 38 layers collapses to 48, and 39 OOMs.
q8_0 KV measured 37 tok/s against 70 on f16 here, so the cache stays f16.

mcp: every wait on the MCP path gets a bound, and a failed reload keeps what it had ...

reload_tools kept dying on prod, and only a restart cured it. Nothing on the
MCP path was bounded: ClientSession was built without read_timeout_seconds, so
the SDK waited forever (anyio.fail_after(None)); _send awaited a future with no
timeout; disconnect() awaited the shielded runner with no timeout; and
disconnect_all() walks the pool one client at a time, so a single wedged client
blocked every reload that came after it. A cancelled caller only cancelled
itself — the runner stayed wedged, which is why it was every reload, not one.

Bounds, in two deliberate layers. The SDK read timeout ends a request nobody
answers *inside* the runner, so the client stays usable; the caller-side bound
is the backstop for an op queued behind a stuck one, and it aborts the runner
rather than leaving it for the next command to queue behind. connect 60s,
list_tools 45s, call_tool 300s (SDK 240s), session read 30s, disconnect 15s,
each overridable, each logging the culprit.

disconnect() and disconnect_all() no longer swallow CancelledError — that is
exactly how the reload budget stopped being enforceable.

The registry is now rebuilt in one step: collect every server's tools first,
then commit the whole set in a single synchronous swap. clear_external_tools()
ran *before* the rebuild, so a rebuild that failed halfway left the model
holding MCP tool names that resolved to nothing. A failed pass now leaves the
previous registrations untouched, and the reload phase carries its own budget
so the tool always returns a report.

Two more defects fell out of the audit. The streamable_http transport was
handed a bare httpx.AsyncClient, whose 5s default read killed the GET stream of
an idle server — six of the nine prod servers are streamable_http. And
register_mcp_tools called log.warning() in a module where log was never
defined, a NameError on precisely the failure path it was meant to report.

Add auth resilience: user cache, retry, and API token fallback ...

- 30-second in-memory _user_cache to avoid hammering gnexus-auth
- _fetch_user_with_retry: one retry after 1.5s sleep on transient failure
- API token fallback when OAuth cookie is present but refresh fails
- Clear cache/locks in test fixture to prevent cross-test pollution
- Fix registry timeout test after lowering default to 90s

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Make Settings immutable (frozen=True) and fix all test mutations ...

- Add frozen=True to SettingsConfigDict in navi/config.py
- Convert model_validator to mode="before" since mode="after" cannot mutate frozen instances
- Replace all field-level monkeypatches in tests with whole-Settings object replacement
- Ensure cross-module settings consistency (content_store, session_files, share_file, content_publish, filesystem)

392 passed, 1 skipped

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Fix ollama_backends / FallbackOllamaBackend issues ...

- registry.py: always use FallbackOllamaBackend (unified backend).
  Enables model priority lists in all deployments, not just multi-server.
- agent.py: add missing think=profile.think_enabled to run() (REST endpoint).
- compressor.py: fix model param type (str → list[str] | str | None).
- fallback.py: harden load_servers_from_file against missing/bad JSON files
  and entries without host. Add clear_blacklists() for manual reset.
- admin.py: add POST /admin/ollama/clear-blacklists endpoint.
- tech_debt_review: document dead stream() methods.
- tests: add tests for single-server fallback, bad file handling,
  missing host skipping, and blacklist clearing.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Add regression tests for content publishing and LLM timeouts

Add context provider registry all accessor

Bootstrap test suite — Phase 1 unit tests ...

- docs/testing.md: testing strategy, mock strategy, phase breakdown
- tests/conftest.py: autouse fixture to reset navi.config.settings per test
- tests/conftest_factory.py: FakeLLMBackend, FakeTool, make_profile, make_registry helpers
- tests/unit/core/test_events.py: wire serialization for all 15 event dataclasses
- tests/unit/core/test_compressor.py: should_compress, partition_messages, format_for_summary, compress_context
- tests/unit/core/test_registry.py: ToolRegistry, ProfileRegistry, BackendRegistry
- tests/unit/core/test_context_builder.py: system prompt caching, persona injection, goal anchor, iteration budget
- tests/unit/profiles/test_base.py: Pydantic model coercion, defaults, extra fields
- navi/core/context_builder.py: use module-level `import navi.config` instead of `from navi.config import settings` so tests can swap the singleton

59 tests passing.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>