From 7877a1bcc931628f8c668ec0feec0917c1c31a93 Mon Sep 17 00:00:00 2001
From: William Valentin <william.valentin.info@gmail.com>
Date: Sun, 15 Feb 2026 22:06:30 -0800
Subject: [PATCH] fix(models): retry timeout errors by default

---
 .../analysis/2026-02-16-codebase-audit-report.md     |  1 +
 docs/plans/state.json                                | 12 ++++++++++++
 src/models/retry.test.ts                             |  4 ++--
 src/models/retry.ts                                  |  3 ---
 4 files changed, 15 insertions(+), 5 deletions(-)

diff --git a/docs/plans/analysis/2026-02-16-codebase-audit-report.md b/docs/plans/analysis/2026-02-16-codebase-audit-report.md
index 9fe0fdf..05bee21 100644
--- a/docs/plans/analysis/2026-02-16-codebase-audit-report.md
+++ b/docs/plans/analysis/2026-02-16-codebase-audit-report.md
@@ -17,6 +17,7 @@ Scope: Production-risk-first audit of bugs, code improvements, and feature oppor
 - ✅ F-014 addressed: `ModelRouter.setOnTierChange` now preserves existing listeners instead of replacing them, removing destructive listener-setter behavior.
 - ✅ F-002 addressed: `config.patch` now supports durable persistence via atomic write + backup when daemon has a concrete config path, and response includes `persisted`/`persistError` so UI can distinguish runtime-only vs disk-persisted updates.
 - ◑ F-003 partially addressed: tool execution now has an `AbortSignal` contract and executor triggers abort on timeout; host `shell.exec` and sandbox docker exec now respond to cancellation. Additional high-risk tools still need explicit cancellation coverage for full closure.
+- ✅ F-015 addressed: retry defaults no longer classify timeout-style failures as non-retryable, improving resilience for transient timeout conditions.
 
 ## Executive Summary
 
diff --git a/docs/plans/state.json b/docs/plans/state.json
index ec6c6c6..15da73d 100644
--- a/docs/plans/state.json
+++ b/docs/plans/state.json
@@ -2565,6 +2565,18 @@
         "docs/plans/analysis/2026-02-16-codebase-audit-report.md"
       ],
       "test_status": "pnpm test:run src/tools/executor.test.ts src/tools/builtin/shell.test.ts + pnpm typecheck passing"
+    },
+    "audit-followup-retry-timeout-defaults": {
+      "status": "completed",
+      "date": "2026-02-16",
+      "updated": "2026-02-16",
+      "summary": "Adjusted retry defaults so timeout-style errors are treated as retryable by default, improving resilience for transient provider/network timeouts.",
+      "files_modified": [
+        "src/models/retry.ts",
+        "src/models/retry.test.ts",
+        "docs/plans/analysis/2026-02-16-codebase-audit-report.md"
+      ],
+      "test_status": "pnpm test:run src/models/retry.test.ts + pnpm typecheck passing"
     }
   },
   "overall_progress": {
diff --git a/src/models/retry.test.ts b/src/models/retry.test.ts
index 85f5a79..ee958b8 100644
--- a/src/models/retry.test.ts
+++ b/src/models/retry.test.ts
@@ -8,9 +8,9 @@ describe('isRetryable', () => {
     expect(isRetryable(error, DEFAULT_RETRY_CONFIG.nonRetryablePatterns)).toBe(true);
   });
 
-  it('returns false for timeout errors', () => {
+  it('returns true for timeout errors (transient)', () => {
     const error = new Error('Request timed out after 20000ms');
-    expect(isRetryable(error, DEFAULT_RETRY_CONFIG.nonRetryablePatterns)).toBe(false);
+    expect(isRetryable(error, DEFAULT_RETRY_CONFIG.nonRetryablePatterns)).toBe(true);
   });
 
   it('returns false for authentication errors', () => {
diff --git a/src/models/retry.ts b/src/models/retry.ts
index cb69d7d..5fc2fbc 100644
--- a/src/models/retry.ts
+++ b/src/models/retry.ts
@@ -26,9 +26,6 @@ export const DEFAULT_RETRY_CONFIG: RetryConfig = {
     'context_length_exceeded',
     'content_policy',
     'does not support',
-    'timeout',
-    'timed out',
-    'request aborted',
   ],
 };