feat(llm-pi-ai): route-keyed profiles with per-request resolution and in-place route swaps

providers becomes a dict keyed by provider route, so the composition base
and the llm-pi-ai settings section merge per provider and the route set is
structural; the pre-release array shape and per-profile provider field fail
loud with migration directions. The adapter reads a profiles thunk once per
operation and resolves the credential per stream call (literal apiKey, then
apiKeyEnv through ctx.credentials with an ambient env fallback, then pi-ai's
provider-native discovery), so key, endpoint, and knob changes reach the
next request without restarts. Route-set or captured-retry-policy changes
re-register the same adapter instance in one synchronous section; an invalid
settings snapshot keeps the last good profiles.

pnpm-lock.yaml

@@ -2977,6 +2977,9 @@ importers:
         specifier: ^3.18.0
         version: 3.18.0
     devDependencies:
+      '@deepseek-ai/dsh-credentials':
+        specifier: workspace:^
+        version: link:../../credentials/credentials
       '@deepseek-ai/dsh-invariants':
         specifier: workspace:^
         version: link:../../support/invariants
@@ -2986,6 +2989,9 @@ importers:
       '@deepseek-ai/dsh-llm-deepseek':
         specifier: workspace:^
         version: link:../llm-deepseek
+      '@deepseek-ai/dsh-settings':
+        specifier: workspace:^
+        version: link:../../settings/settings
       '@deepseek-ai/dsh-timeout':
         specifier: workspace:^
         version: link:../../util/timeout