skills.rest
kinetas
kinetasCOMMUNITY
Explore all skills by this author →
GitHub

debugLimit

Set the maximum Debugger AI rejection count in doc/company_state.json.

Updated May 21, 2026
Search Tags:#code review#debugging#state management#ai workflow#loop prevention#iteration limits
One-click install
npx skills add https://github.com/kinetas/harness_engineering --skill debuglimit
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: debugLimit
Source: https://github.com/kinetas/harness_engineering/tree/main/debugLimit
Command: npx skills add https://github.com/kinetas/harness_engineering --skill debuglimit

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The debugLimit skill prevents infinite reject/redo loops between Debugger AI and Developer AI by capping how many times Debugger AI can reject work before escalation.

Core Features & Use Cases

  • Argument-driven limit setting: Updates the rejection allowance using an optional numeric argument provided with the command.
  • Input validation safeguards: Enforces a minimum of 1 and warns that values above 5 increase infinite loop risk.
  • State persistence & user confirmation flow: Writes the updated value to doc/company_state.json and reports the change back to the user.

Use case: When your AI workflow repeatedly fails the same code-review or debugging stage, use debugLimit to keep iterations bounded and ensure the system escalates to Boss AI.

Quick Start

Invoke /debugLimit 3 to set the Debugger AI rejection allowance to 3.

Frequently Asked Questions about debugLimit

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I stop infinite debugging loops between Developer AI and Debugger AI?▼

To stop infinite debugging loops, you can cap the maximum number of times Debugger AI rejects Developer AI output. This prevents endless iteration cycles during code review and ensures the workflow escalates to Boss AI.

How do I set a maximum rejection limit for AI code review iterations?▼

You can set a maximum rejection limit by invoking the command with a numeric argument, such as /debugLimit 3. This updates the rejection allowance and writes the validated value to the persistent state file.

What is the safe range for setting AI debugging rejection limits?▼

The safe range for AI debugging rejection limits requires a minimum value of 1. Values above 5 are allowed but will trigger a warning because they significantly increase the risk of infinite loop stalls.

Where is the AI debugging rejection count state persisted?▼

The AI debugging rejection count state is persisted in the doc/company_state.json file. Writing the updated limit to this JSON file ensures deterministic escalation behavior across your workflow.

When should I enforce iteration limits on AI code improvement workflows?▼

You should enforce iteration limits on AI code improvement workflows when repeated debugging rejections stall progress. Bounding iterations keeps the debugging cycle finite and forces deterministic escalation to Boss AI.

Related Skills

Explore More Software Engineering Skills→
kinetaskinetas

startCompany

Create Harness Engineering state files and initialize AI orchestration.

Community
Intermediate
masharrattmasharratt
14

cfn-intervention-system

Detect human intervention requirements in CFN Loop automation workflows.

Community
Advanced
kinetaskinetas

autoReport

Toggle Collector AI's auto report collection state for report.md maintenance.

Community
Basic
oozoofrogoozoofrog
1

ralph-loop

Run bounded Codex loops with completion criteria and maximum iterations.

Community
Advanced