[BUG] 'insights' command fails - ValidationException: output_config.effort not permitted

Preflight Checklist

• [x] I have searched existing issues and this hasn't been reported yet
• [x] This is a single bug report (please file separate reports for different bugs)
• [x] I am using the latest version of Claude Code

What's Wrong?

/insights command fails - ValidationException: output_config.effort not permitted

Description

The /insights command fails to generate AI-powered content sections, returning AWS Bedrock ValidationException errors because it's passing an unsupported parameter (output_config.effort) to the API.

Error

400: ValidationException when calling the Converse operation:
The model returned the following errors: output_config.effort: Extra inputs are not permitted

Impact

• All AI-generated sections fail: at_a_glance, project_areas, interaction_style, what_works, friction_analysis, suggestions, on_the_horizon, fun_ending
• Basic statistics and charts work correctly
• Generated HTML report contains only numeric data, no narrative insights

Environment

• Claude Code Version: 2.1.119
• Platform: macOS 26.4.1 (ARM64)
• Model: Claude Sonnet 4.5

Actual Behavior

• Basic statistics are collected successfully (sessions count, messages, tool usage, charts)
• All AI-generated narrative sections fail with identical ValidationException
• HTML report is generated but all insight sections remain empty

Root Cause

The insights feature appears to be passing output_config.effort parameter that AWS Bedrock API no longer accepts (or never accepted in the current API version).

Additional Context

• Report HTML file location: ~/.claude/usage-data/report.html
• Basic analytics (stats, tool usage charts, time-of-day distribution) work correctly
• Only AI-generated narrative sections fail
• Error appears consistently across all insight categories

Similar, yet different bugs:

• https://github.com/anthropics/claude-code/issues/38224
• https://github.com/anthropics/claude-code/issues/37854
• https://github.com/anthropics/claude-code/issues/32227

What Should Happen?

Expected Behavior

All sections should contain AI-generated analysis of usage patterns.

Error Messages/Logs

## Sample Error Response


{
  "Error": {
    "Message": "The model returned the following errors: output_config.effort: Extra inputs are not permitted",
    "Code": "ValidationException"
  },
  "ResponseMetadata": {
    "HTTPStatusCode": 400,
    "MaxAttemptsReached": true,
    "RetryAttempts": 0
  }
}

Steps to Reproduce

Steps to Reproduce

1. Run /insights command
2. Check generated report at ~/.claude/usage-data/report.html
3. All insight sections are empty with ValidationException errors

Claude Model

Sonnet (default)

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.119 (Claude Code)

Platform

AWS Bedrock

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_