[Bug] Sub-agent recursive delete via typo'd path destroys codebase on NTFS

Resolved 💬 3 comments Opened May 18, 2026 by RSSRINDIA Closed Jun 18, 2026

Bug Description
CRITICAL — codebase destroyed A Phase 2 sub-agent I spawned for row #290 MarksheetReprint ran a recursive delete against a typo'd path. On NTFS that resolved to the real codebase path and wiped most of D:\CampusConnect\. I verified the destruction directly with ls: Gone: - D:\CampusConnect\MVC\ — entire legacy MVC source tree - D:\CampusConnect\Core\UniversityManagementSystem.Contracts\ — directory removed - D:\CampusConnect\Core\UniversityManagementSystem.Api\ — directory empty - D:\CampusConnect\Core\UniversityManagementSystem.Bff\ — directory empty - D:\CampusConnect\Core\UniversityManagementSystem.Application\ — directory empty - D:\CampusConnect\Core\UniversityManagementSystem.Infrastructure\ — directory empty except a ResultProcessing shell subdir - D:\CampusConnect\Core\ums-spa\src\ — directory removed - D:\CampusConnect\scripts\ — directory removed - All root .md files including MIGRATION_INDEX.md, MIGRATION_EXECUTION_PROMPT.md, MIGRATION_GUIDE_GENERATION_PROMPT.md, CLAUDE.md, bold_reports.md - sql.txt who to recover my codebase

Environment Info

  • Platform: win32
  • Terminal: windows-terminal
  • Version: 2.1.112
  • Feedback ID: 2a12808c-bf4f-4f0d-b539-10b46f6ba2f7

Errors

[{"error":"MaxFileReadTokenExceededError: File content (45829 tokens) exceeds maximum allowed tokens (25000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at QyK (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4980:12515)\n    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)\n    at async dyK (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4984:2638)\n    at async Object.call (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4989:7236)\n    at async UJY (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4439:8320)","timestamp":"2026-05-18T02:12:48.002Z"},{"error":"Error: ENOENT: no such file or directory, unlink 'C:\\Users\\RSSRSOFT\\AppData\\Local\\Temp\\claude\\D--CampusConnect\\0f643918-7194-469b-8404-075114c27eb2\\tasks\\ab1603a51d5d03611.output'\n    at async unlink (node:internal/fs/promises:1065:10)\n    at async file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:8166:340","timestamp":"2026-05-18T02:29:59.217Z"},{"error":"Error: ENOENT: no such file or directory, unlink 'C:\\Users\\RSSRSOFT\\AppData\\Local\\Temp\\claude\\D--CampusConnect\\0f643918-7194-469b-8404-075114c27eb2\\tasks\\aeea605df89a87163.output'\n    at async unlink (node:internal/fs/promises:1065:10)\n    at async file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:8166:340","timestamp":"2026-05-18T02:30:32.581Z"},{"error":"MaxFileReadTokenExceededError: File content (81001 tokens) exceeds maximum allowed tokens (25000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at QyK (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4980:12515)\n    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)\n    at async dyK (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4984:2638)\n    at async Object.call (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4989:7236)\n    at async UJY (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4439:8320)","timestamp":"2026-05-18T02:30:46.111Z"},{"error":"MaxFileReadTokenExceededError: File content (37424 tokens) exceeds maximum allowed tokens (25000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at QyK (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4980:12515)\n    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)\n    at async dyK (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4984:2638)\n    at async Object.call (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4989:7236)\n    at async UJY (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthropic-ai/claude-code/cli.js:4439:8320)","timestamp":"2026-05-18T02:31:11.157Z"},{"error":"MaxFileReadTokenExceededError: File content (37424 tokens) exceeds maximum allowed tokens (25000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n    at QyK (file:///C:/Users/RSSRSOFT/AppData/Roaming/npm/node_modules/@anthro…

Note: Content was truncated.

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗