v2.4.1-sidecar-routing: route per-agent flags to llama-sidecar + tool gap fix

Batch 3c: when an agent has llama_extra_args in AGENTS.md, provider.ts
routes inference through LLAMA_SIDECAR_URL instead of LLAMA_SWAP_URL.
X-Agent-Flags header built from the agent's flags. Boot-time guard
refuses to start if any agent has llama_extra_args but LLAMA_SIDECAR_URL
is unset. PrefixFingerprint gains a route field (swap/sidecar) for
per-turn visibility. 9 provider tests.

AGENTS.md tool gap: all agents (except Prompt Builder) were missing 8
tools that were added after the original tool lists were written:
request_read_access, view_truncated_output, ask_user_input, git_status,
get_blast_radius, get_hot_files, get_middleware, get_routes. The missing
request_read_access caused silent "permission denied" when reading files
outside the project root.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

apps/server/src/services/inference/provider.ts

@@ -1,37 +1,84 @@
 import { createOpenAICompatible } from '@ai-sdk/openai-compatible';
 import type { LanguageModel } from 'ai';
 
-// TODO: When per-agent llama-server flag overrides are added, route them
-// through validateExtraArgs (./llama-args-validator.ts) first.
-
 // v1.13.1-A: AI SDK provider against llama-swap. baseURL is threaded from
 // config.LLAMA_SWAP_URL at call time (not module-load) so tests can stub the
 // upstream without touching env vars. No apiKey — llama-swap is unauth in our
 // Tailscale topology and exposing it over the public internet is gated by
 // Authelia at the Caddy layer, not by API keys.
+//
+// v2.4.1-sidecar: when the agent has llama_extra_args, route through
+// llama-sidecar instead. A fresh provider is created per call (not cached)
+// because the X-Agent-Flags header varies per agent. The llama-swap path
+// stays cached since it has no per-request headers.
 
-const cache = new Map<string, ReturnType<typeof createOpenAICompatible>>();
+const swapCache = new Map<string, ReturnType<typeof createOpenAICompatible>>();
 
-function getProvider(baseURL: string): ReturnType<typeof createOpenAICompatible> {
-  let provider = cache.get(baseURL);
+function getSwapProvider(baseURL: string): ReturnType<typeof createOpenAICompatible> {
+  let provider = swapCache.get(baseURL);
   if (!provider) {
     provider = createOpenAICompatible({
       name: 'llama-swap',
       baseURL: baseURL.endsWith('/v1') ? baseURL : `${baseURL}/v1`,
-      // v1.13.7: @ai-sdk/openai-compatible defaults includeUsage=false, which
-      // omits `stream_options.include_usage` from the request body. Without
-      // it, llama.cpp / llama-swap never emits the trailing usage block, so
-      // `result.usage` resolves with inputTokens=outputTokens=undefined and
-      // tokens_used / ctx_used land as NULL in every messages row. Setting
-      // true here re-enables the per-stream usage payload across all models
-      // served via the llama-swap provider.
       includeUsage: true,
     });
-    cache.set(baseURL, provider);
+    swapCache.set(baseURL, provider);
   }
   return provider;
 }
 
-export function upstreamModel(baseURL: string, modelId: string): LanguageModel {
-  return getProvider(baseURL).chatModel(modelId);
+function sidecarProvider(
+  baseURL: string,
+  flags: string[],
+): ReturnType<typeof createOpenAICompatible> {
+  return createOpenAICompatible({
+    name: 'llama-sidecar',
+    baseURL: baseURL.endsWith('/v1') ? baseURL : `${baseURL}/v1`,
+    includeUsage: true,
+    headers: {
+      'X-Agent-Flags': flags.join(' '),
+    },
+  });
+}
+
+export type InferenceRoute = 'swap' | 'sidecar';
+
+export interface RoutingInfo {
+  route: InferenceRoute;
+  flags: string[] | null;
+}
+
+interface AgentLike {
+  llama_extra_args: string[] | null;
+}
+
+interface ConfigLike {
+  LLAMA_SWAP_URL: string;
+  LLAMA_SIDECAR_URL?: string;
+}
+
+export function resolveRoute(agent: AgentLike | null): RoutingInfo {
+  const flags = agent?.llama_extra_args;
+  if (flags && flags.length > 0) {
+    return { route: 'sidecar', flags };
+  }
+  return { route: 'swap', flags: null };
+}
+
+export function upstreamModel(
+  config: ConfigLike,
+  modelId: string,
+  agent?: AgentLike | null,
+): LanguageModel {
+  const { route, flags } = resolveRoute(agent ?? null);
+  if (route === 'sidecar') {
+    const url = config.LLAMA_SIDECAR_URL;
+    if (!url) {
+      throw new Error(
+        `Agent has llama_extra_args but LLAMA_SIDECAR_URL is not set`,
+      );
+    }
+    return sidecarProvider(url, flags!).chatModel(modelId);
+  }
+  return getSwapProvider(config.LLAMA_SWAP_URL).chatModel(modelId);
 }