@askalf/dario 4.8.46 → 4.8.48

package/dist/cc-template.d.ts

@@ -291,14 +291,31 @@ export declare function parseEffortSuffix(model: string): {
  *               accepted by all and still routes to the subscription pool
  *               (verified: representative-claim=five_hour on Opus + Sonnet).
  *               Set --effort=xhigh / DARIO_EFFORT=xhigh for Opus's extra tier.)
+ *               EXCEPT fable: real CC's print-mode default for fable is
+ *               'high' (live capture 2026-06-09, CC v2.1.170 — the same
+ *               capture got 'xhigh' on opus and 'high' on fable), so the
+ *               unset-flag default mirrors that per-family. An explicit
+ *               flag still pins (subject to the fable clamp below).
  *   'low' / 'medium' / 'high' / 'xhigh' / 'max' → pin to that value
  *   'ultracode' → 'xhigh' (CC's ultracode mode; xhigh on the wire)
  *   'client' → extract from `clientBody.output_config.effort` (normalized
- *              for the wire); fall back to 'max' if absent/non-string
+ *              for the wire); fall back to the per-family default if
+ *              absent/non-string
+ *
+ * FABLE CLAMP (2026-06-09, live replay bisect on the deployed proxy):
+ * fable-5 SOFT-REFUSES `max` and `xhigh` — 200 with stop_reason "refusal"
+ * and empty content on every prompt, not a 400 — while `high` answers
+ * normally (byte-identical request bodies, only output_config.effort
+ * mutated). Empirical matrix on claude-fable-5:
+ *   high  → answers      xhigh → refusal      max → refusal
+ * So on the fable family any resolved 'max'/'xhigh' (from --effort /
+ * DARIO_EFFORT pins, ultracode, or client passthrough) is clamped to
+ * 'high', the strongest level fable accepts. Operators who pin
+ * --effort=max for Opus's sake keep that on every other family.
  *
  * Exported for tests.
  */
-export declare function resolveEffort(flag: EffortValue | undefined, clientBody: Record<string, unknown>): string;
+export declare function resolveEffort(flag: EffortValue | undefined, clientBody: Record<string, unknown>, model?: string): string;
 /**
  * Returns true if the given model accepts `thinking: { type: "adaptive" }`.
  *

package/dist/cc-template.js

@@ -985,24 +985,44 @@ function normalizeEffortForWire(effort) {
  *               accepted by all and still routes to the subscription pool
  *               (verified: representative-claim=five_hour on Opus + Sonnet).
  *               Set --effort=xhigh / DARIO_EFFORT=xhigh for Opus's extra tier.)
+ *               EXCEPT fable: real CC's print-mode default for fable is
+ *               'high' (live capture 2026-06-09, CC v2.1.170 — the same
+ *               capture got 'xhigh' on opus and 'high' on fable), so the
+ *               unset-flag default mirrors that per-family. An explicit
+ *               flag still pins (subject to the fable clamp below).
  *   'low' / 'medium' / 'high' / 'xhigh' / 'max' → pin to that value
  *   'ultracode' → 'xhigh' (CC's ultracode mode; xhigh on the wire)
  *   'client' → extract from `clientBody.output_config.effort` (normalized
- *              for the wire); fall back to 'max' if absent/non-string
+ *              for the wire); fall back to the per-family default if
+ *              absent/non-string
+ *
+ * FABLE CLAMP (2026-06-09, live replay bisect on the deployed proxy):
+ * fable-5 SOFT-REFUSES `max` and `xhigh` — 200 with stop_reason "refusal"
+ * and empty content on every prompt, not a 400 — while `high` answers
+ * normally (byte-identical request bodies, only output_config.effort
+ * mutated). Empirical matrix on claude-fable-5:
+ *   high  → answers      xhigh → refusal      max → refusal
+ * So on the fable family any resolved 'max'/'xhigh' (from --effort /
+ * DARIO_EFFORT pins, ultracode, or client passthrough) is clamped to
+ * 'high', the strongest level fable accepts. Operators who pin
+ * --effort=max for Opus's sake keep that on every other family.
  *
  * Exported for tests.
  */
-export function resolveEffort(flag, clientBody) {
+export function resolveEffort(flag, clientBody, model) {
+    const isFable = (model ?? '').toLowerCase().includes('fable');
+    const familyDefault = isFable ? 'high' : 'max';
+    const clamp = (effort) => isFable && (effort === 'max' || effort === 'xhigh') ? 'high' : effort;
     if (flag === undefined)
-        return 'max';
+        return familyDefault;
     if (flag === 'client') {
         const clientOC = clientBody.output_config;
         const clientEffort = clientOC?.effort;
         if (typeof clientEffort === 'string' && clientEffort.length > 0)
-            return normalizeEffortForWire(clientEffort);
-        return 'max';
+            return clamp(normalizeEffortForWire(clientEffort));
+        return familyDefault;
     }
-    return normalizeEffortForWire(flag);
+    return clamp(normalizeEffortForWire(flag));
 }
 /**
  * Returns true if the given model accepts `thinking: { type: "adaptive" }`.
@@ -1477,7 +1497,7 @@ export function buildCCRequest(clientBody, billingTag, cacheControl, identity, o
             }
         }
         if (!skip || !skip.has('output_config')) {
-            ccRequest.output_config = { effort: resolveEffort(opts.effort, clientBody) };
+            ccRequest.output_config = { effort: resolveEffort(opts.effort, clientBody, model) };
         }
     }
     ccRequest.stream = stream;

package/dist/proxy.d.ts

@@ -17,6 +17,18 @@ export declare function parseProviderPrefix(model: string): {
     provider: 'openai' | 'claude';
     model: string;
 } | null;
+/**
+ * Fable-5 (2026-06) — real CC appends `fallback-credit-2026-06-01` to the
+ * beta set on fable requests ONLY (live captures 2026-06-09, CC v2.1.170:
+ * the fable request carries it, the opus request does not). Subscription
+ * traffic on fable WITHOUT it is soft-refused upstream — every request
+ * returns 200 with `stop_reason: "refusal"` and empty content, while the
+ * same request on opus/sonnet answers normally. Mirror CC exactly: append
+ * for the fable family, leave every other family's set untouched.
+ * Exported for tests.
+ */
+export declare const FABLE_FALLBACK_CREDIT_BETA = "fallback-credit-2026-06-01";
+export declare function betaForModel(base: string, model: string | null | undefined): string;
 export declare const ORCHESTRATION_TAG_NAMES: string[];
 /**
  * Build the regex list that actually strips orchestration tags.

package/dist/proxy.js

@@ -190,6 +190,24 @@ const BILLABLE_BETA_PREFIXES = [
 function filterBillableBetas(betas) {
     return betas.split(',').map(b => b.trim()).filter(b => b.length > 0 && !BILLABLE_BETA_PREFIXES.some(p => b.startsWith(p))).join(',');
 }
+/**
+ * Fable-5 (2026-06) — real CC appends `fallback-credit-2026-06-01` to the
+ * beta set on fable requests ONLY (live captures 2026-06-09, CC v2.1.170:
+ * the fable request carries it, the opus request does not). Subscription
+ * traffic on fable WITHOUT it is soft-refused upstream — every request
+ * returns 200 with `stop_reason: "refusal"` and empty content, while the
+ * same request on opus/sonnet answers normally. Mirror CC exactly: append
+ * for the fable family, leave every other family's set untouched.
+ * Exported for tests.
+ */
+export const FABLE_FALLBACK_CREDIT_BETA = 'fallback-credit-2026-06-01';
+export function betaForModel(base, model) {
+    if (!model || !model.toLowerCase().includes('fable'))
+        return base;
+    if (base.split(',').includes(FABLE_FALLBACK_CREDIT_BETA))
+        return base;
+    return base ? `${base},${FABLE_FALLBACK_CREDIT_BETA}` : FABLE_FALLBACK_CREDIT_BETA;
+}
 // Orchestration tags injected by agents (Aider, Cursor, OpenCode, etc.)
 // that confuse Claude when passed through. Strip before forwarding.
 export const ORCHESTRATION_TAG_NAMES = [
@@ -1562,7 +1580,9 @@ export async function startProxy(opts = {}) {
                 // skip context-1m entirely (dario#36).
                 const acctKey = poolAccount?.alias ?? ACCOUNT_KEY_SINGLE;
                 const skipContext1m = context1mUnavailable.has(acctKey);
-                beta = skipContext1m ? betaWithoutContext1m : betaBase;
+                // Fable requires its fallback-credit beta or upstream soft-refuses
+                // every request (see betaForModel) — model-conditional, like real CC.
+                beta = betaForModel(skipContext1m ? betaWithoutContext1m : betaBase, requestModel);
                 if (clientBeta) {
                     const baseSet = new Set(beta.split(','));
                     const filtered = filterBillableBetas(clientBeta)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@askalf/dario",
-  "version": "4.8.46",
+  "version": "4.8.48",
   "description": "Use your Claude Pro/Max subscription in any tool — Cursor, Cline, Aider, the Agent SDK, your scripts — at subscription pricing, not per-token API bills. One local Anthropic + OpenAI-compatible endpoint.",
   "type": "module",
   "bin": {