Maintainer ruling on the review's merge-semantics warning: per-model
reasoning fields belong to the settings document, not cordis.yml entry
config (the shipped composition mounts the adapter dormant), so the
recursive-merge delete gap is a documented posture rather than a tracked
fix; the Known Limitations entry now states the assumption instead of
pointing at the closed#2003.
Review findings on #1977, each verified before acting: the 2026-08-03
declared-provider-catalog note is updated in place and cross-linked both
ways now that reasoningEfforts/compat reopened half of its rejected
alternative; resolveModelCompat inherits the catalog entry's compat only
while the resolved api still is the entry's own, so a route-level api
repoint no longer merges another protocol's shape as a completions base;
the off-with-value promise gains a request-boundary test proving pi-ai
reads thinkingLevelMap.off when the reasoning option is absent (and the
catalog-level test name stops overclaiming); the cannot-stop-thinking
wording narrows to what is actually enforced (no Off offered, explicit
Off refused — an effortless request goes out bare); the z.const(null)
comment attributes null passthrough to schemastery's nullable
short-circuit; the baseten drift-gate claim names its verification
source; and the layered-merge delete gap for dict keys is documented
under Known Limitations with the atomic-leaf follow-up in #2003.
A route's modelOverrides dict customizes individual installed-catalog
models — key = catalog model id, value = the same fields a models entry
takes — while the rest of the catalog keeps serving, which a models list
cannot express because declaring one replaces the served set. An
override becomes the catalog entry's configuration and resolves through
the existing entry path, so capacities, reasoningEfforts, compat, and
request-default semantics are identical to a models entry's. Unlike
Pi's config layer, which ignores unknown ids, every override that lands
nowhere is refused at the write that produced it: beside a models list,
on a hand-declared route, naming a model the catalog does not describe,
or smuggling an id through the schema's unknown-key tolerance.
A model entry's reasoningEfforts dict declares its selectable thinking
levels — key = offered level, value = the wire spelling dispatch sends;
only off may leave the value empty (supported, send nothing). false
strips reasoning from a catalog model; every level is materialized
explicitly into pi-ai's thinkingLevelMap so nobody has to know pi-ai's
asymmetric absent-key defaulting. compat.thinkingFormat and
compat.supportsReasoningEffort become configurable on the route and per
model (model > route > catalog entry > pi-ai's URL-derived guess),
openai-completions only, so a private gateway speaking the DeepSeek
reasoning dialect no longer depends on its URL being recognizable.
Record-typed drift gates pin both enums to pi-ai's, and an unserviceable
declaration is refused at the write that produced it, naming route,
model, and level.
Code already treats $DSH_HOME/.env as ordinary launch environment and stores managed credentials in .credentials.yaml, but public docs still described the old store, old precedence, removed literal adapter keys, and the deleted TUI. That directed users to the wrong file and overstated the supported configuration surface.
Update the existing English and Chinese owners in place, document inherited > managed > project > user credential resolution, and record the loadLayeredEnv export. Regenerate only pairing records and the source-line catalog; add no new section or site route.
`resolveModel` validated the profile's reasoning level against the exact
model and threw when it did not fit. That call builds the model catalog,
and the catalog build catches per PROVIDER — so one mis-set field took the
whole provider out of every picker behind a single error row, hiding even
the models that do support the level. Measured: `anthropic` set to `max`
threw for six of its eight models.
Describing what a model can do now reports an unusable profile level as no
default rather than throwing; the request path still refuses it, which is
where a bad configuration belongs. The existing spec asserted the old
throw and now asserts both halves of that split.
Known gap, left deliberately: a model that cannot take the route's level
still fails its first request while the picker shows 「Default」 for it,
because the request path keeps using the profile level as the fallback.
Reaching that needs a hand-written `settings.yaml` — the Models page no
longer writes the field — and the error names the model and the level, so
selecting a supported level is a way out. Closing it properly means giving
`AgentOptions` a `reasoningEffort` so compositions without a model picker
keep an entry point, then dropping the provider-scoped field altogether;
that is its own change.
Carries two edits beyond conflict resolution, both forced by what master
brought in:
- `CustomProviderCard`: master added front-end key validation and a
component-level `keyValue` (already trimmed) while still writing
`apiKeyEnv` unconditionally. Kept this branch's blank-key rule and its
committed-profile retry gate, and adopted master's single `keyValue` so
the component has one spelling of the key rather than two.
- `docs/user/guide/providers`: master merged #1810, whose default-model
section still taught overriding the `api-gateway` row in
`$DSH_HOME/config.yaml` — the behavior this branch replaced. Rewritten
for the settings section the picker now writes, plus the review fix from
#1810 replacing the colloquial 挂着 in the opener.
A row's stored profile could not tell a hand-declared gateway from a
shipped provider whose models someone narrowed — both look identical from
outside the adapter — so the Models page had no way to mark the routes a
deployment added itself.
The directory entry now carries `declared`, answered by the owning adapter
against its own installed catalog, and the page renders a Custom tag from
it. Absence stays "this adapter draws no such distinction" rather than
"shipped", so a route no adapter claims is labelled neither way.
Also records the default-route work's Agent Note and the e2e evidence for
all three changes: the composer switch writing the section, and the Models
page declaring a route with its own reasoning effort.
A configuration surface never holds a stored secret — it edits a redacted
descriptor — so once a key is saved, the draft it sends carries the route and
the endpoint and no credential at all. The interrogation went out
unauthenticated and the endpoint's 401 came back as "check the API key",
pointing at the one thing that was fine.
A named route now supplies its own credential, resolved exactly as a request
to it would be. A key typed into the form still wins: it is the one under
test, and may be the replacement for the stored one that is failing.
Resolution is a callback the probe invokes past the catalog short-circuit and
the protocol check, so a route answered from the installed registry costs no
credential lookup — and cannot fail over a credential the question never
needed.
llm.discoverModels was reachable from any declared trusted host. The
method takes a caller-supplied baseURL and makes the host issue a GET to
it, then reports the status or the parsed body — so on a LAN deployment
an anonymous caller had a probe for whatever the host can reach and the
browser cannot, plus a path that carries a draft credential. The
PRIVILEGED_METHODS doc already states the rule this broke: trustedHosts
is a DNS-rebinding fence, not authentication, so the configuration plane
stays loopback-same-origin. It is in that set now, asserted both against
the hand-built fence and over real HTTP beside the catalog reads that
deliberately stay reachable.
supportsDiscovery and listModelDiscoveryNamespaces are gone. The field
was required on the wire and read by nobody: its own contract said a
surface should offer the action "instead of naming an adapter family it
would have to hardcode", while the surface hardcodes llm-pi-ai in two
places and gates the button on whether there is anything to probe. Its
shape did not fit the second caller either — the create card has no row
to read a per-row field from. Keeping a required field alive for a
consumer that may never arrive costs every producer and fixture a value
nobody consults, which is exactly how the fixtures drifted. The registry
that fed it had no other production consumer, so registration and
disposal are now observed through the offer itself.
The Agent Note claimed the key is never logged, which the wire schema
beside it already contradicts, and predated both the provider field and
the catalog-answer path. The two new public types pointed at core.md
without a type-equiv block or manifest entry, so the generated service
catalog named documentation that did not exist.
pi-ai resolves a request's apiKey override only through a provider that
declares an api-key method: resolveProviderAuth short-circuits to that
method when the override is present, and otherwise falls through to the
credential store and then to ambient discovery. A provider with no
api-key method at all therefore resolves to nothing, and the request
fails with "Provider is not configured" before any network I/O.
Two routes hit that. openai-codex ships OAuth alone, so moving off the
/compat dispatch broke a profile that names a key for it — the old path
handed the token straight to the provider. And a catalog route naming an
api was being rebuilt with the harness's own auth, so `openai: {api:
openai-completions}` stopped reading OPENAI_API_KEY, contradicting the
documented promise that omitting a credential keeps provider-native
discovery.
Auth is now one decision for both constructions. A catalog route keeps
its installed provider's auth, through an api override too: which
environment a provider reads belongs to the provider, not to the wire
format its models speak. A catalog provider with no api-key method gets
the harness method beside its own, but only when the profile names a
credential — a keyless codex profile keeps the honest refusal, since
this adapter holds no OAuth store to resolve through.
Materialization now spreads the installed entry instead of enumerating
the result, so a Model field this package does not model survives a
pi-ai upgrade; headers went missing from an nvidia route exactly that
way once already. providerInfo reports the configured displayName, which
also joins the registration facts so a rename re-registers rather than
leaving the old label in every selector. A refused registration swap
gets its own diagnostic naming the route, matching the directory swap
beside it.
The README documented endpoint interrogation this layer does not
implement, and still described unknown providers as kept-last-good after
they became legal declarations refused at the write point. The Agent
Note claimed per-model reasoning configurability the schema never had,
required capacities the route now defaults, and stated an apiKey
override that short-circuits unconditionally.
Three defects surfaced while driving the Models page.
A hand-declared model needed an explicit contextWindow and maxTokens,
but a provider listing usually returns ids and nothing else — so the
page happily wrote a profile the adapter then rejected, which took the
whole namespace down silently. Capacities now fall back to the route's
`defaultContextWindow` (262,144) and `defaultMaxTokens` (32,768). Both
are guesses by construction, which is why they are route fields a
deployment corrects once rather than constants buried in the adapter;
the fallback sizes the model and never becomes a per-request cap.
That silent failure was the second defect. A schema-valid profile the
adapter could not serve was stored and only rejected later, disabling
every route in the namespace with nothing said. `dsh-settings` gains an
optional `validate` on registration — a check for what a schema cannot
express — and `llm-pi-ai` refuses an unserviceable section at the write
that produced it. A stored section that fails keeps the namespace's last
good value, as a schema failure already did, so an externally edited
document still cannot strand the owner. The plugin's own last-good
fallback goes with it: nothing reaching it can fail any more.
Third, a model with no reasoning metadata advertised the single level
`off`, which pi-ai translates to *omitting* the reasoning option — the
same request naming no effort produces. Selecting it disabled nothing,
so a provider whose default is to think kept thinking with `off` shown
as selected. Such a model now reports no reasoning capability at all,
which is the seam's way of saying the control is unavailable, and the
per-model `reasoning` flag is gone: without a thinkingLevelMap to spell
levels it could only invent them.
The protocol table narrows to the three a hand-declared route reaches
today, most-reached first so a surface offering a choice defaults to the
one gateways actually speak.
Review found four defects in the declared-provider work.
`PiAiAdapter` reused one `Models` collection and mutated it whenever the
configuration changed. `Models.streamSimple()` resolves its provider
lazily — when the stream is first consumed, which is after the adapter
awaits the route's credential — so a configuration change landing in
that window let an in-flight request finish under a configuration it
never resolved against, or fail on a provider that no longer existed.
Each resolution now produces an immutable snapshot and every operation
captures one before its first await, which is what makes the seam's
per-step freeze (`llm.prepareCall()`) hold end to end: switching models
mid-reply takes effect on the next step, never inside the one in flight.
`defaultMaxTokens` was materialized from the catalog's `Model.maxTokens`.
The two answer different questions: pi-ai requires that field as the
model's output capability, while the seam's is a cap the deployment
chose to send on requests naming none, so every request had started
carrying a number nobody picked. Only an explicitly configured cap
reaches the seam now.
The configurable-provider directory was refreshed by disposing its
registration and making a new one. A candidate set the registry refuses
— a profile keyed `deepseek-official`, which llm-deepseek declares —
left the whole directory withdrawn and the Models page empty, silently,
because the settings callback contains the failure. The seam's
registration handle now carries `replace()` with the same
validate-first atomicity `registerAdapter` has.
The protocol table offered every pi-ai streaming API, including four
whose authentication a profile cannot express: Bedrock signs with SigV4
over AWS credentials and a region, Vertex needs a project, a location,
and ADC, Azure needs provider environment plus an api-version, and Codex
uses OAuth. Offering them handed back routes that cannot authenticate.
Catalog routes still reach them through their own provider.
A pi-ai route had to name an installed catalog provider, served that
catalog's models verbatim, and could override only the endpoint. An
OpenAI-compatible gateway, a self-hosted server, or a model newer than
the pinned pi-ai release was therefore unreachable, and a stale context
window could not be corrected without upgrading the package.
A route is now a declaration whose defaults come from the installed
catalog. `catalog.ts` merges that catalog under the profile's own model
entries, `provider.ts` builds the pi-ai Provider (reusing the catalog
provider when the route keeps its protocol, so implementations this
package cannot reconstruct keep working), and the adapter serves every
operation from one `createModels()` collection. That also retires the
`@earendil-works/pi-ai/compat` import, which pi-ai documents as a
temporary entry point it deletes with its ModelManager migration.
Credentials stay on the harness seam: the resolved key rides the request
as pi-ai's highest-priority auth override, so `Models` holds no
credential store and a named-but-missing reference still fails loud
instead of falling back to an unrelated ambient key.
A model's configured maxTokens now reaches the seam as defaultMaxTokens.
The remaining P1 from the #939 review, plus the P2 it shares a mechanism with.
Nothing carried a version, so two tabs editing one namespace silently
overwrote each other — reproduced as tab B's `reasoning` lost to tab A's
older draft. The seam's per-namespace write queue orders writes; it cannot
tell a fresh writer from one replaying a snapshot a predecessor superseded.
Each namespace now carries a monotonic `revision` over its RAW section. A
write may send `expectedRevision`, checked at the FRONT of the queue (not at
call time, which would race the very predecessor it guards against); a
mismatch rejects with `SettingsConflictError` → `settings-conflict` on the
wire, carrying both revisions. The editor captures the revision it opened at
and, on conflict, asks the user to reopen rather than replaying its snapshot.
The same counter fixes the missing broadcast. `settings/updated` is gated on
the resolved value — correct for consumers, wrong for configuration surfaces:
storing an override equal to the composition base leaves the resolved value
alone while changing what the document says (the field is now overridden, not
inherited) and moving every open editor's revision. `settings/document-updated
(ns, revision)` fires on any raw-section change, in-process or external, and
`host/settings-changed` now rides it.
That event also closes the stale model picker: editing a provider's `models`
changes no route, so `llm/adapters-updated` never fired and an open picker
kept serving the old catalog. A change to an exposed provider namespace now
emits `host/models-changed` too — that namespace holds the catalog.
Docs: both sides of the five touched README pairs, a type-equiv block for
`SettingsPathOp`, and an Agent Note recording what the plane exposes and who
may overwrite what. The deferred wire-redaction gaps (secrets behind
union/intersection/transform, `.default(...)` in the served envelope, schema
text in rejection messages, `new Function` rehydration, pi-ai's `headers`) are
recorded as TODO(settings-wire-redaction) and in Known Limitations rather than
half-fixed.
The policy home's resolve() now stamps readDenyPaths, so every consumer
that pins the resolved shape (bash-sandbox hand-off, tool-fs stamps)
carries it, and three uncovered branches gained real tests: landlock
reporting partial enforcement for a denial it cannot express, the
policy's default under programmatic construction, and both ambient
credential paths in llm-deepseek without a mounted seam.
ctx.llm gains 'llm/adapters-updated' — a payload-free registry notification
emitted at every topology commit point (adapter routes registering or
disposing, directory entries appearing or withdrawing) with contained
observers and INVARIANT rethrow — plus registerConfigurableProviders/
listConfigurableProviders, the directory of routes an adapter plugin can
activate through configuration. llm-deepseek declares deepseek-official
(whole llm-deepseek section as profile); llm-pi-ai declares the full
installed catalog under providers.<provider> even while dormant, so the
web settings surface can offer every provider before any route exists.
The invariant companion asserts the registry stays readable at each
notification.
An empty or omitted providers dict is now the valid dormant posture: the
adapter mounts with zero routes and no catalog entries, registers routes
the moment the llm-pi-ai settings section supplies profiles, and drops
them when it empties. The TUI demo mounts the adapter bare, so adding an
openai/anthropic provider is purely a settings.yaml (or, next PR, web
form) operation with per-request apiKeyEnv credential resolution.
New credentials data-structure page (type-equiv manifested), group README,
rewritten llm-deepseek/llm-pi-ai READMEs (dynamic configuration, dict
profiles, credential chain), capability-seams/service-role registration,
Agent Note (bilingual), demo compositions mounting settings-local +
credentials-local with no inline key plumbing, installSettingsSection
consumer helper on the settings seam (deduplicating both adapters' wiring),
jscpd symmetry markers for the provider twins, runtime-closure additions for
python/sdk-runtime, and doc-budget ceilings AGENTS.md 1750→1755 /
packages/README.md 850→865 for the structural one-line group rows.
A well-formed provider stream that ends with finish_reason stop and zero
content blocks previously became a successful empty assistant message: the
turn completed silently, and drivers like goal-session counted the no-op
round. Both adapters now map that degenerate completion to a finish
{kind:'error'} with the new canonical EMPTY_RESPONSE code from dsh-llm, and
dsh-llm-retry adds the code to its default retryable set, so the existing
closed-step recovery path retries it and fails loud once the budget is
exhausted.
Covered by adapter unit tests, an llm-retry default-policy test, and a new
authored keyless ACP snapshot (empty-response-retry) with a deterministic
1 ms zero-jitter retry overlay.
Pass each resolved catalog model capacity into pi-ai stream conversion so the upstream full-message classifier can recognize provider-specific, silent, and length-stop overflow signals. Retain the harness text fallback for legacy provider wording and cover the catalog-resolution path with a mock-provider regression.
Turtle