auth: openai_responses provider for Codex subscription (Tier 2)

Add the OpenAI Responses API wire dialect (openai_responses APIStyle +
OpenAIResponsesConfig), a request serializer (instructions/input items,
flat function tools, reasoning, store:false, include encrypted reasoning),
and a streaming state machine over the typed response.* SSE events mapping
to the shared block model. Wire dispatch, buildProviderConfig, the C-ABI
free switch, and the reasoning selectors. Ship a commented Codex example in
the default config.

Implemented from the Responses API docs + the open-source Codex client and
covered by fixture tests; live verification against a ChatGPT plan is still
required (documented).

1 files changed, 24 insertions, 0 deletions

diff --git a/libpanto/src/config.zig b/libpanto/src/config.zig
index 09aaec8..c48cefc 100644
--- a/libpanto/src/config.zig
+++ b/libpanto/src/config.zig
@@ -24,6 +24,9 @@ const Io = std.Io;
 pub const APIStyle = enum {
     openai_chat,
     anthropic_messages,
+    /// OpenAI Responses API (`/responses`). Used by the ChatGPT-subscription
+    /// Codex backend, whose request/stream shape differs from Chat Completions.
+    openai_responses,
 };
 
 /// A single HTTP request header (name/value). Used for provider
@@ -133,10 +136,25 @@ pub const AnthropicMessagesConfig = struct {
     extra_headers: []const Header = &.{},
 };
 
+/// OpenAI Responses API config. Same transport knobs as `openai_chat`, but a
+/// distinct request body (`input` items, `reasoning`, `store`, `include`) and
+/// streaming event shape. Auth-derived headers (`chatgpt-account-id`) and the
+/// static Codex identity headers (`OpenAI-Beta`, `originator`) ride on
+/// `extra_headers`.
+pub const OpenAIResponsesConfig = struct {
+    api_key: []const u8,
+    base_url: []const u8,
+    model: []const u8,
+    reasoning: ReasoningEffort = .default,
+    max_tokens: u32 = 64_000,
+    extra_headers: []const Header = &.{},
+};
+
 /// Per-provider transport/auth/model configuration. Tagged by `APIStyle`.
 pub const ProviderConfig = union(APIStyle) {
     openai_chat: OpenAIChatConfig,
     anthropic_messages: AnthropicMessagesConfig,
+    openai_responses: OpenAIResponsesConfig,
 
     pub fn style(self: ProviderConfig) APIStyle {
         return @as(APIStyle, self);
@@ -163,6 +181,12 @@ pub const ProviderConfig = union(APIStyle) {
                 .thinking_budget_tokens = c.thinking_budget_tokens,
                 .thinking_interleaved = c.thinking_interleaved,
             },
+            .openai_responses => |c| .{
+                .api_style = .openai_responses,
+                .base_url = c.base_url,
+                .model = c.model,
+                .reasoning = c.reasoning,
+            },
         };
     }
 };