Add OpenAI platform: live model fetching with enrichment (provider 1)

The 4th registry row: id "openai", OPENAI_API_KEY env > auth.json
"openai" scope (login picker/paste/validation all registry-generic —
zero TUI changes needed, pinned by the picker test), base
https://api.openai.com/v1 with KIGI_OPENAI_BASE_URL override, Responses
dialect via the new PlatformWireApi spec field, enrichment-backed
metadata (wire_serves_metadata=false).

OpenAI's GET /v1/models returns bare ids and is polluted with
tts/whisper/embeddings entries: the listing is restricted to
enrichment-known TOOL-CALLING models (review caught that membership
alone admitted models.dev-known embeddings models, which would 400 on
every agentic request; dropped ids are debug-logged for launch-day
diagnosability). Context windows, effort menus, display names, and
thinking capability come from the enrichment pipeline — wiremock e2e
pins the full contract: polluted live listing + models.dev →
one Responses-backed chat model with a 400k documented context window.

Responses max-effort wiring (closes the P0c-1 debt): canonical effort
rides a CreateResponseWrapper sidecar and patch_reasoning_effort writes
it onto the serialized body at both send sites (all seven levels pinned,
xhigh/max distinct, summary preserved); normalize_effort_echo drops
echoes async-openai's typed enum cannot represent at both the non-stream
and SSE parse seams; the dead typed to_responses_api converter is
deleted. Kimi/moonshot stay byte-identical (ChatCompletions untouched,
wire_api maps to the same default; kimi wire tests green).

kimi-import now recognizes ANY registry platform host as built-in
(was hardcoded moonshot), covering openai and future rows.

crates/codegen/kigi-sampling-types/src/conversation.rs

@@ -2158,7 +2158,10 @@ impl From<&ConversationRequest> for rs::CreateResponse {
             prompt_cache_key: None,
             prompt_cache_retention: None,
             reasoning: Some(rs::Reasoning {
-                effort: req.reasoning_effort.map(|e| e.to_responses_api()),
+                // Effort is written onto the serialized body by
+                // `patch_reasoning_effort` (the typed enum cannot spell
+                // `max`); the wrapper carries the canonical value.
+                effort: None,
                 summary: Some(rs::ReasoningSummary::Concise),
             }),
             safety_identifier: None,
@@ -5191,6 +5194,10 @@ mod tests {
 
     #[test]
     fn test_responses_request_carries_reasoning_effort_nested() {
+        // Wire contract since the Max split: the typed conversion leaves
+        // reasoning.effort UNSET (async-openai's enum cannot spell `max`);
+        // `patch_reasoning_effort` writes the canonical token onto the
+        // serialized body — every level, `xhigh` and `max` distinct.
         for (variant, expected) in [
             (crate::ReasoningEffort::None, "none"),
             (crate::ReasoningEffort::Minimal, "minimal"),
@@ -5198,6 +5205,7 @@ mod tests {
             (crate::ReasoningEffort::Medium, "medium"),
             (crate::ReasoningEffort::High, "high"),
             (crate::ReasoningEffort::Xhigh, "xhigh"),
+            (crate::ReasoningEffort::Max, "max"),
         ] {
             let req = ConversationRequest {
                 reasoning_effort: Some(variant),
@@ -5205,15 +5213,56 @@ mod tests {
                     .with_model("test")
             };
             let resp: crate::rs::CreateResponse = (&req).into();
-            let json = serde_json::to_value(&resp).unwrap();
+            let mut json = serde_json::to_value(&resp).unwrap();
+            assert_eq!(
+                json.pointer("/reasoning/effort"),
+                None,
+                "typed conversion must leave effort unset ({variant:?})"
+            );
+            crate::patch_reasoning_effort(&mut json, req.reasoning_effort);
             assert_eq!(
                 json.pointer("/reasoning/effort").and_then(|v| v.as_str()),
                 Some(expected),
-                "{variant:?} should serialize as reasoning.effort={expected:?}; got: {json:#}",
+                "{variant:?} should be patched as reasoning.effort={expected:?}; got: {json:#}",
+            );
+            assert_eq!(
+                json.pointer("/reasoning/summary").and_then(|v| v.as_str()),
+                Some("concise"),
+                "the patch must not clobber reasoning.summary"
             );
         }
     }
 
+    #[test]
+    fn normalize_effort_echo_drops_only_unrepresentable_tokens() {
+        // `max` echo (bare response shape) is dropped so typed parsing
+        // succeeds; known tokens pass through; stream-event envelopes are
+        // handled via /response/reasoning.
+        let mut bare = serde_json::json!({ "reasoning": { "effort": "max", "summary": "c" } });
+        crate::normalize_effort_echo(&mut bare);
+        assert_eq!(bare.pointer("/reasoning/effort"), None);
+        assert_eq!(
+            bare.pointer("/reasoning/summary").and_then(|v| v.as_str()),
+            Some("c")
+        );
+
+        let mut known = serde_json::json!({ "reasoning": { "effort": "high" } });
+        crate::normalize_effort_echo(&mut known);
+        assert_eq!(
+            known.pointer("/reasoning/effort").and_then(|v| v.as_str()),
+            Some("high"),
+            "representable echoes must pass through"
+        );
+
+        let mut event = serde_json::json!({ "response": { "reasoning": { "effort": "max" } } });
+        crate::normalize_effort_echo(&mut event);
+        assert_eq!(event.pointer("/response/reasoning/effort"), None);
+
+        let mut absent = serde_json::json!({ "output": [] });
+        crate::normalize_effort_echo(&mut absent);
+        assert_eq!(absent, serde_json::json!({ "output": [] }));
+    }
+
     #[test]
     fn test_responses_request_omits_effort_when_unset() {
         let req =

crates/codegen/kigi-sampling-types/src/types.rs

@@ -794,34 +794,11 @@ pub enum ReasoningEffort {
 }
 
 impl ReasoningEffort {
-    pub fn to_responses_api(self) -> crate::rs::ReasoningEffort {
-        match self {
-            Self::None => crate::rs::ReasoningEffort::None,
-            Self::Minimal => crate::rs::ReasoningEffort::Minimal,
-            Self::Low => crate::rs::ReasoningEffort::Low,
-            Self::Medium => crate::rs::ReasoningEffort::Medium,
-            Self::High => crate::rs::ReasoningEffort::High,
-            Self::Xhigh => crate::rs::ReasoningEffort::Xhigh,
-            // INTERIM: async-openai (0.33 pinned; upstream through 0.41
-            // still lacks it) has no `Max` variant. The Responses backend
-            // gains true `max` via a post-serialize body patch in the OpenAI
-            // provider cycle; until then Max downgrades to xhigh — reachable
-            // via `[model.*] reasoning_effort = "max"` on a responses-backend
-            // model, hence the loud warn.
-            Self::Max => {
-                tracing::warn!(
-                    "reasoning effort `max` downgraded to `xhigh` on the \
-                     Responses backend (typed wire enum has no max yet)"
-                );
-                crate::rs::ReasoningEffort::Xhigh
-            }
-        }
-    }
-
-    /// Inverse of [`to_responses_api`](Self::to_responses_api): the effort the
-    /// Responses API echoes back on `response.reasoning.effort`. (`max`
-    /// echoes cannot reach here — async-openai's enum has no such variant;
-    /// the OpenAI provider cycle handles them before typed parsing.)
+    /// The canonical effort behind a typed Responses-API echo
+    /// (`response.reasoning.effort`). `max` echoes never reach the typed
+    /// enum — [`normalize_effort_echo`] drops them pre-parse (async-openai
+    /// has no such variant); the request direction writes the wire string
+    /// via [`patch_reasoning_effort`].
     pub fn from_responses_api(effort: crate::rs::ReasoningEffort) -> Self {
         match effort {
             crate::rs::ReasoningEffort::None => Self::None,
@@ -911,6 +888,52 @@ where
     }))
 }
 
+/// Write the canonical effort onto a serialized Responses request body
+/// (`body.reasoning.effort`). This is the ONLY place effort reaches the
+/// Responses wire: the typed `rs::ReasoningEffort` tops out at `xhigh`, so
+/// `max` must be written post-serialize. A `None` effort leaves the body
+/// untouched (the provider default applies).
+pub fn patch_reasoning_effort(body: &mut Value, effort: Option<ReasoningEffort>) {
+    let Some(effort) = effort else { return };
+    let Some(obj) = body.as_object_mut() else {
+        return;
+    };
+    let reasoning = obj
+        .entry("reasoning")
+        .or_insert_with(|| Value::Object(serde_json::Map::new()));
+    if let Some(reasoning) = reasoning.as_object_mut() {
+        reasoning.insert(
+            "effort".to_string(),
+            Value::String(effort.as_str().to_string()),
+        );
+    }
+}
+
+/// Neutralize a `reasoning.effort` echo the typed `rs` enum cannot parse
+/// (`max`): remove it so response deserialization succeeds. The turn's
+/// canonical effort lives in the session sampling config regardless; only
+/// the per-item echo metadata is dropped, recorded as not-echoed.
+/// (Ceiling: async-openai lacks a Max variant; delete this when it grows
+/// one.) Handles both bare response bodies (`/reasoning/effort`) and
+/// stream-event envelopes (`/response/reasoning/effort`).
+pub fn normalize_effort_echo(value: &mut Value) {
+    for path in ["/reasoning", "/response/reasoning"] {
+        if let Some(reasoning) = value.pointer_mut(path).and_then(|v| v.as_object_mut())
+            && let Some(effort) = reasoning.get("effort").and_then(|v| v.as_str())
+            && crate::rs::ReasoningEffort::deserialize(serde_json::Value::String(
+                effort.to_string(),
+            ))
+            .is_err()
+        {
+            tracing::debug!(
+                effort,
+                "reasoning.effort echo unrepresentable in the typed enum; dropping"
+            );
+            reasoning.remove("effort");
+        }
+    }
+}
+
 pub const REASONING_EFFORT_META_KEY: &str = "reasoningEffort";
 pub const SUPPORTS_REASONING_EFFORT_META_KEY: &str = "supportsReasoningEffort";
 
@@ -1126,6 +1149,12 @@ pub struct CreateResponseWrapper {
     /// The inner Responses API request.
     pub inner: crate::rs::CreateResponse,
 
+    /// Canonical reasoning effort for this request. The typed
+    /// `inner.reasoning.effort` stays `None` (async-openai's enum cannot
+    /// spell `max`); [`patch_reasoning_effort`] writes this value onto the
+    /// serialized body just before send.
+    pub reasoning_effort: Option<ReasoningEffort>,
+
     /// Custom header: conversation ID for tracking.
     pub x_kigi_conv_id: Option<String>,
 
@@ -1152,6 +1181,7 @@ impl CreateResponseWrapper {
     pub fn new(inner: crate::rs::CreateResponse) -> Self {
         Self {
             inner,
+            reasoning_effort: None,
             x_kigi_conv_id: None,
             x_kigi_req_id: None,
             x_kigi_session_id: None,