# Read Ue Logs

> Read and filter Unreal Engine log output from disk for any UE project. Auto-detects the project from the nearest *.uproject, merges entries across all concurrently-active log files (editor session, standalone clients), and surfaces only recently-active sessions by default. Use when diagnosing a silent failure after a repro, checking startup warnings, inspecting what UE printed for a recent gamepl…

- **Type:** Skill
- **Install:** `agentstack add skill-osseous-skills-read-ue-logs`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [osseous](https://agentstack.voostack.com/s/osseous)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [osseous](https://github.com/osseous)
- **Source:** https://github.com/osseous/skills/tree/main/skills/read-ue-logs

## Install

```sh
agentstack add skill-osseous-skills-read-ue-logs
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# read-ue-logs

## Quick start

From the project root, run:

```
powershell -NoProfile -File .claude/skills/read-ue-logs/scripts/read-logs.ps1 -Tail 50
```

By default this scans `Saved/Logs/.log` plus any `Saved/Logs/_N.log` modified in the last 10 minutes, merges them by timestamp, and prints the last 50 lines. The script auto-locates the project root and project name from the nearest `*.uproject`, so the working directory does not need to be the repo root. Use `pwsh -File ...` instead of `powershell -File ...` if PowerShell 7+ is installed (both editions are supported).

> **Running UE automation tests?** Run them through a test MCP server — [`osseous/ue-headless-mcp`](https://github.com/osseous/ue-headless-mcp) (preferred) or [`remiphilippe/mcp-unreal`](https://github.com/remiphilippe/mcp-unreal) — (`run_tests` / `run_visual_tests`) and take the first read of the result from its `get_test_log`. Reach for *this* skill when one log isn't enough — merging concurrent editor + standalone-client logs into one timeline, filtering by category/verbosity/source, or grepping across a whole session.

## Workflows

### Diagnose a failure after a repro

```
# Recent errors only
powershell -NoProfile -File .claude/skills/read-ue-logs/scripts/read-logs.ps1 -Verbosity Error -Tail 100

# Blueprint-specific output
powershell -NoProfile -File .claude/skills/read-ue-logs/scripts/read-logs.ps1 -Category LogBlueprint -Tail 100

# Hunt for a specific keyword (regex)
powershell -NoProfile -File .claude/skills/read-ue-logs/scripts/read-logs.ps1 -Search "MyActor|Spawn"
```

### Analyze a multi-client test run

When multiple instances run (editor + standalone clients), UE writes `.log`, `_2.log`, `_3.log` etc. Default behavior merges all of them and prefixes each line with `[]`.

```
# All clients, last 10 min, merged and prefixed
powershell -NoProfile -File .claude/skills/read-ue-logs/scripts/read-logs.ps1 -Tail 200

# Isolate one client (substitute your actual project name)
powershell -NoProfile -File .claude/skills/read-ue-logs/scripts/read-logs.ps1 -Source MyProject_2
```

### Programmatic consumption

```
powershell -NoProfile -File .claude/skills/read-ue-logs/scripts/read-logs.ps1 -Format json -Tail 50
```

Each line is one JSON object: `{source, line, timestamp, thread, category, verbosity, message}`.

## Advanced features

- `-RecentMinutes ` — change the auto-discovery window (default 10). Pass a large value to merge older sessions; if nothing matches the window, the script falls back to the single most-recent log so it never returns empty.
- `-LogFile ` — point at one specific log file, bypassing auto-discovery. Pass `-LogFile all` to include rotated `-backup-*.log` files too.
- `-ProjectName ` — override the auto-detected project name for unusual setups (renamed logs, multi-uproject monorepos).
- Exit codes: `0` on success (even with zero matches), `1` only on a missing log directory or script error.
- The editor does not need to be running — the script reads files on disk.

See [README.md](README.md) for the full flag reference, the multi-file model, and notes on extending the parser.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [osseous](https://github.com/osseous)
- **Source:** [osseous/skills](https://github.com/osseous/skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-osseous-skills-read-ue-logs
- Seller: https://agentstack.voostack.com/s/osseous
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.