fix(sidecar): sanitize JSONL session files before resume (strips empty text blocks with cache_control)

sidecar/claude-runner.ts

@@ -8,7 +8,8 @@ import { execSync } from 'child_process';
 import { existsSync } from 'fs';
 import { join } from 'path';
 import { homedir } from 'os';
-import { query, type Query } from '@anthropic-ai/claude-agent-sdk';
+import { query, listSessions, type Query } from '@anthropic-ai/claude-agent-sdk';
+import { readFileSync, writeFileSync, readdirSync, statSync } from 'fs';
 
 const rl = createInterface({ input: stdin });
 
@@ -121,12 +122,17 @@ async function handleQuery(msg: QueryMessage) {
     if (resumeMode === 'continue') {
       continueOpt = true;
       log(`Session ${sessionId}: continuing most recent session`);
+      // Sanitize the most recent session file before SDK reads it
+      sanitizeSessionFiles(cwd || process.cwd());
     } else if (resumeMode === 'resume' && resumeSessionId) {
       resumeOpt = resumeSessionId;
       log(`Session ${sessionId}: resuming SDK session ${resumeSessionId}`);
+      // Sanitize the specific session file
+      sanitizeSessionFile(cwd || process.cwd(), resumeSessionId);
     } else if (resumeSessionId && !resumeMode) {
       // Legacy: direct resumeSessionId without resumeMode
       resumeOpt = resumeSessionId;
+      sanitizeSessionFile(cwd || process.cwd(), resumeSessionId);
     }
 
     const q = query({
@@ -238,5 +244,108 @@ if (claudePath) {
   log('WARNING: Claude CLI not found — agent sessions will fail');
 }
 
+// ── Session sanitizer ──────────────────────────────────────────────────────
+// Fixes "cache_control cannot be set for empty text blocks" API error on resume.
+// The SDK stores empty text blocks with cache_control during streaming;
+// the API rejects them on replay. We strip them before the SDK reads the file.
+
+function encodeCwd(cwdPath: string): string {
+  return cwdPath.replace(/[^a-zA-Z0-9]/g, '-');
+}
+
+function sanitizeSessionFile(cwdPath: string, sdkSessionId: string): void {
+  const encoded = encodeCwd(cwdPath);
+  const filePath = join(homedir(), '.claude', 'projects', encoded, `${sdkSessionId}.jsonl`);
+  try {
+    sanitizeJsonlFile(filePath);
+  } catch (err) {
+    log(`sanitize: could not clean ${filePath}: ${err}`);
+  }
+}
+
+function sanitizeSessionFiles(cwdPath: string): void {
+  // For 'continue' mode, sanitize the most recent session
+  try {
+    const sessions = listSessions({ dir: cwdPath, limit: 1 });
+    if (sessions && sessions.length > 0) {
+      sanitizeSessionFile(cwdPath, (sessions[0] as Record<string, unknown>).sessionId as string);
+    }
+  } catch (err) {
+    log(`sanitize: listSessions failed: ${err}`);
+    // Fallback: try to find and sanitize the most recent .jsonl file
+    const encoded = encodeCwd(cwdPath);
+    const dir = join(homedir(), '.claude', 'projects', encoded);
+    try {
+      const files = readdirSync(dir).filter(f => f.endsWith('.jsonl'));
+      if (files.length > 0) {
+        // Sort by mtime descending
+        files.sort((a, b) => {
+          try {
+            return statSync(join(dir, b)).mtimeMs - statSync(join(dir, a)).mtimeMs;
+          } catch { return 0; }
+        });
+        sanitizeJsonlFile(join(dir, files[0]));
+      }
+    } catch { /* dir doesn't exist */ }
+  }
+}
+
+function sanitizeJsonlFile(filePath: string): void {
+  if (!existsSync(filePath)) return;
+
+  const content = readFileSync(filePath, 'utf-8');
+  const lines = content.split('\n');
+  let modified = false;
+
+  const cleaned = lines.map(line => {
+    if (!line.trim()) return line;
+    try {
+      const obj = JSON.parse(line);
+      if (cleanContentBlocks(obj)) {
+        modified = true;
+        return JSON.stringify(obj);
+      }
+    } catch { /* skip unparseable lines */ }
+    return line;
+  });
+
+  if (modified) {
+    writeFileSync(filePath, cleaned.join('\n'));
+    log(`sanitize: cleaned empty text blocks from ${filePath}`);
+  }
+}
+
+function cleanContentBlocks(obj: Record<string, unknown>): boolean {
+  let changed = false;
+
+  // Check message.content array
+  const msg = obj.message as Record<string, unknown> | undefined;
+  if (msg?.content && Array.isArray(msg.content)) {
+    const before = msg.content.length;
+    msg.content = (msg.content as Array<Record<string, unknown>>).filter(block => {
+      // Remove empty text blocks (with or without cache_control)
+      if (block.type === 'text' && (!block.text || !(block.text as string).trim())) {
+        return false;
+      }
+      return true;
+    });
+    if (msg.content.length !== before) changed = true;
+  }
+
+  // Also check top-level content array (some formats)
+  if (obj.content && Array.isArray(obj.content)) {
+    const before = (obj.content as unknown[]).length;
+    obj.content = (obj.content as Array<Record<string, unknown>>).filter(block => {
+      if (block.type === 'text' && (!block.text || !(block.text as string).trim())) {
+        return false;
+      }
+      return true;
+    });
+    if ((obj.content as unknown[]).length !== before) changed = true;
+  }
+
+  return changed;
+}
+
 log('Sidecar started');
 send({ type: 'ready' });