diff --git a/.github/workflows/claude-review.yml b/.github/workflows/claude-review.yml
index d55d585..7bdde86 100644
--- a/.github/workflows/claude-review.yml
+++ b/.github/workflows/claude-review.yml
@@ -2,7 +2,7 @@ name: Claude Code Review
 
 on:
   pull_request:
-    types: [opened, ready_for_review]
+    types: [opened, synchronize, ready_for_review]
   issue_comment:
     types: [created]
 
@@ -16,6 +16,7 @@ jobs:
   review-pr:
     if: github.event_name == 'pull_request'
     runs-on: ubuntu-latest
+    timeout-minutes: 15
 
     steps:
       - name: Checkout
@@ -25,6 +26,10 @@ jobs:
         uses: anthropics/claude-code-action@v1
         with:
           anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
+          claude_args: >-
+            --allowedTools
+            "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*),Bash(gh pr diff:*),Bash(gh pr view:*)"
+            --max-turns 10
           prompt: |
             REPO: ${{ github.repository }}
             PR NUMBER: ${{ github.event.pull_request.number }}
@@ -39,6 +44,7 @@ jobs:
       github.event.issue.pull_request &&
       contains(github.event.comment.body, '@claude')
     runs-on: ubuntu-latest
+    timeout-minutes: 15
 
     steps:
       - name: Checkout
@@ -48,3 +54,7 @@ jobs:
         uses: anthropics/claude-code-action@v1
         with:
           anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
+          claude_args: >-
+            --allowedTools
+            "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*),Bash(gh pr diff:*),Bash(gh pr view:*)"
+            --max-turns 10
diff --git a/README.md b/README.md
index 2af9527..08efeb1 100644
--- a/README.md
+++ b/README.md
@@ -35,7 +35,7 @@ node dist/cli/index.js --help
 - `linkledger index-rebuild [--json]`
 - `linkledger worker [--limit N] [--max-attempts N] [--base-backoff-ms N] [--json]`
 
-## Ingestion adapters (M2)
+## Ingestion adapters
 
 - `article`: HTML extraction and chunking.
 - `x`: oEmbed-based extraction with fallback to article adapter.
@@ -47,7 +47,7 @@ node dist/cli/index.js --help
 
 Retryable adapter failures are requeued with exponential backoff in `worker`.
 
-In M3, successful ingest also creates enrichment artifacts and moves items to `enriched`.
+Successful ingest also creates enrichment artifacts (summary, key claims) and moves items to `enriched` status.
 
 ## Database path
 
diff --git a/src/adapters/reddit-adapter.ts b/src/adapters/reddit-adapter.ts
index 1f213cb..0f95a88 100644
--- a/src/adapters/reddit-adapter.ts
+++ b/src/adapters/reddit-adapter.ts
@@ -7,6 +7,9 @@ import type { AdapterParseResult, SourceAdapter } from './source-adapter.js';
 
 const MAX_POST_CHARS = 1800;
 const MAX_COMMENT_CHARS = 900;
+// limit=8 in the fetch URL over-fetches to buffer for non-comment children
+// (e.g. kind:'more' stubs) that get filtered out, ensuring we still collect
+// up to MAX_COMMENTS actual t1 comments.
 const MAX_COMMENTS = 5;
 
 interface RedditListingChild {
@@ -59,15 +62,6 @@ const toPublishedAt = (value: unknown): string | undefined => {
   return new Date(value * 1000).toISOString();
 };
 
-const isRedditPageUrl = (value: string): boolean => {
-  try {
-    const parsed = new URL(value);
-    return isRedditHost(parsed.hostname.toLowerCase());
-  } catch {
-    return false;
-  }
-};
-
 export class RedditAdapter implements SourceAdapter {
   supports(url: string): boolean {
     return this.detectType(url) === 'reddit';
@@ -109,13 +103,15 @@ export class RedditAdapter implements SourceAdapter {
     const subreddit = compact(toText(postData.subreddit));
     const selfText = compact(toText(postData.selftext));
     const linkedUrl = compact(toText(postData.url));
+    let isLinkedReddit = false;
+    try { isLinkedReddit = isRedditHost(new URL(linkedUrl).hostname.toLowerCase()); } catch { /* malformed URL */ }
 
     const postParts = [
       title ? `Title: ${title}` : '',
       subreddit ? `Subreddit: r/${subreddit}` : '',
       author ? `Author: u/${author}` : '',
       selfText ? `Body:\n${selfText}` : '',
-      linkedUrl && !isRedditPageUrl(linkedUrl) ? `Linked URL: ${linkedUrl}` : ''
+      linkedUrl && !isLinkedReddit ? `Linked URL: ${linkedUrl}` : ''
     ].filter(Boolean);
 
     const chunks: Array<{ text: string; tokenCount: number }> = [];
diff --git a/test/unit/reddit-adapter.test.ts b/test/unit/reddit-adapter.test.ts
index aaff48b..c81ef74 100644
--- a/test/unit/reddit-adapter.test.ts
+++ b/test/unit/reddit-adapter.test.ts
@@ -2,6 +2,7 @@ import assert from 'node:assert/strict';
 import { readFileSync } from 'node:fs';
 import path from 'node:path';
 import test from 'node:test';
+import { AppError } from '../../src/lib/errors.js';
 import { RedditAdapter } from '../../src/adapters/reddit-adapter.js';
 
 const fixture = (name: string): string =>
@@ -57,3 +58,28 @@ test('RedditAdapter marks upstream failures as retryable for 5xx', async () => {
     globalThis.fetch = originalFetch;
   }
 });
+
+test('RedditAdapter marks 404 as non-retryable', async () => {
+  const adapter = new RedditAdapter();
+  const originalFetch = globalThis.fetch;
+
+  globalThis.fetch = async () =>
+    new Response('not found', {
+      status: 404,
+      headers: { 'content-type': 'text/plain' }
+    });
+
+  try {
+    await assert.rejects(
+      () => adapter.fetchAndParse({ url: 'https://www.reddit.com/comments/abc123' }),
+      (error: unknown) => {
+        assert.ok(error instanceof AppError);
+        assert.equal(error.code, 'FETCH_FAILED');
+        assert.equal(error.retryable, false);
+        return true;
+      }
+    );
+  } finally {
+    globalThis.fetch = originalFetch;
+  }
+});