Skip to content
Back to skills

Hayba Debug Level

ASecurity

Use when an Unreal level has placement, streaming, physics, or performance symptoms that need a measured diagnosis and verified fix.

  • 13 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added October 3, 2026
ai-agentsgodebuggingperformance

Security analysis

A100/100

Scanned October 4, 2026

npx -y skills add zajalist/hayba --skill hayba-debug-level --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Hayba Debug Level?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Hayba Debug Level
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/zajalist-hayba-debug-level/badge)](https://www.skillsdirectory.com/skills/zajalist-hayba-debug-level)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: hayba-debug-level
description: Use when an Unreal level has placement, streaming, physics, or performance symptoms that need a measured diagnosis and verified fix.
metadata:
  category: qa
  tags: debugging, physics, streaming, performance
---

# Diagnose a level symptom

## When to use

Use for a concrete symptom such as floating objects, overlaps, missing generated content, low frame rate, or a streaming gap. Record the camera, active map, and reproduction steps so a later readback is comparable.

## Tool routing

- Start with `hayba_check_ue_status`. Gather a baseline through `editor_get_performance_stats`, `editor_stream_log`, and `scene_export` in hierarchical mode. Use `wp_get_cells` when World Partition is relevant.
- For placement symptoms, run `scene_validate_physics`; inspect affected actors with `actor_list` and `placement_validate` before considering `actor_transform` or `actor_set_visibility`. A position that looks unusual may be an intentional user adjustment.
- For generated content, use the hayba-pcg-build skill and a fresh cook readback. For performance, change one suspected cause at a time, then repeat the same measurement conditions.
- For an accepted map repair, read back the changed actor or generated output, call `level_save` for the exact active map, and inspect saved, dirty, and verified status. If persistence cannot be checked safely, report the repair as unsaved and route it to hayba-world-save-integrity.

## Evidence

Give the original symptom, baseline values or findings, the smallest confirmed cause, the edit made, and the after measurement. Run `validator_run` after scene edits. If measurements are unavailable, leave the diagnosis as a hypothesis rather than calling the level fixed.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.