PostToolUse hooks should fire for failed tool calls (not just successful ones)

Summary

This is a follow-up to #6371, which was auto-closed due to inactivity. The feature request remains relevant.

Current behavior: PostToolUse hooks only fire when tools complete successfully. Failed tool calls (e.g., Bash commands with non-zero exit codes) skip the hook entirely.

Requested behavior: PostToolUse should fire for all tool completions, with success: false and failure details when the tool fails.

Refined Proposal (from #6371 discussion)

Instead of introducing a separate PostToolFailure hook, modify PostToolUse to:

1. Fire for all tool completions (success or failure)
2. Include "success": true for successful calls (current behavior)
3. Include "success": false plus a failure_details object for failed calls:

{
  "session_id": "abc123...",
  "hook_event_name": "PostToolUse",
  "tool_name": "Bash",
  "tool_input": {
    "command": "ls nonexistent-file.txt"
  },
  "success": false,
  "failure_details": {
    "exit_code": 2,
    "stdout": "",
    "stderr": "ls: cannot access 'nonexistent-file.txt': No such file or directory",
    "error_type": "CommandFailed"
  }
}

The existing tool_response.success field already exists in the schema, suggesting this may have been the original design intent.

Use Cases

• Analytics & logging: Track all tool usage including failures to improve MCP server instructions and tool schemas
• Contextual error feedback: Hook can provide Claude with better context (e.g., "file not found - here are files in current directory")
• Automated remediation hints: Parse stderr and suggest fixes (e.g., git push --set-upstream origin branch-name)
• Cleanup after failures: Remove temp files/directories created before a multi-step operation failed

Original Issue

See #6371 for full reproduction steps, debug output evidence, and detailed discussion. Credit to the original author and commenters (@Twizzes, @coygeek, @richardkmichael) for the thorough analysis.