base44-troubleshooter Skill for Claude: What It Does and How to Install It

Makes Claude debug a production Base44 app by reading its function logs and failed workflow runs.

Installs: 10.8K; on skills.sh
Stars: 92; on GitHub
Files: 1; besides SKILL.md
Updated: Nov 2026; last repo commit

This is Base44's own skill for production debugging. It tells Claude to run npx base44 logs and npx base44 workflows commands, starting with a live --follow stream if you can trigger the failure, and with failed workflow runs if the complaint is a scheduled job that did not run.

It needs the Base44 CLI signed in, or an --app-id. The folder holds SKILL.md and one reference file for the logs command, and bundles no scripts.

Good for

Base44 developers chasing errors in backend functions
Anyone whose scheduled job or automation stopped running
People who want to watch function output live while they trigger a failure

Install base44-troubleshooter

Claude Code

Copy this prompt.

Install the base44-troubleshooter skill in Claude Code. Run this command:

/plugin marketplace add base44/skills
/plugin install base44@base44-skills

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Pick one route, not several. The plugin route installs 12 skills: base44-cli, base44-sdk, base44-troubleshooter and 9 base44-platform skills (counted from plugin.json). The skills CLI route names base44-cli next to this skill because SKILL.md links to base44-cli reference files for the workflow commands. The README's Antigravity package includes base44-troubleshooter.

From the project README. Source.

What's in the skill

When it loads

Troubleshoot production issues using backend function logs and workflow run history. Use when investigating app errors, debugging function calls, diagnosing why a scheduled job or automation failed, or diagnosing production problems in Base44 apps.

Sections in SKILL.md

Prerequisites
Available Commands
Logs are not read-after-write
Troubleshooting Flow
1. Watch it happen — --follow
2. Ask whether it was a scheduled run, not a request
3. Check Recent Errors
4. Drill Into a Specific Function
5. Inspect a Time Range
6. Analyze the Logs
Reading an empty result

Key facts

Author: Base44 (vendor's own repo)
Repo: base44/skills
License: MIT (LICENSE file; README, marketplace.json and plugin.json agree)
Installs on skills.sh: 10,756 (Nov 2, 2026)
GitHub stars: 92 for the whole repo (Nov 2, 2026)
Last repo commit: Nov 1, 2026
Audits on skills.sh: Gen Agent Trust Hub: pass, Socket: pass, Snyk: pass
What it includes: 2 files including SKILL.md: 1 markdown reference file and no scripts
Part of: Plugins base44 and base44-sandbox
Needs: Base44 CLI login; links to base44-cli reference files
Status: Beta, per the README

What it includes

Runs npx base44 whoami and asks you to log in if it fails
Runs code · Auth check
Reads base44/.app.jsonc or takes --app-id to pick the app
Files · App context
Streams function logs live while you trigger the failure
Where the realtime stream is available lines arrive in under a second. Otherwise the CLI warns and polls, with lines lagging about 20 to 30 seconds.
Try: “Watch the logs for my processOrder function while I run it”
Runs code · logs --follow
Lists failed workflow runs and workflow health for scheduled jobs
Runs `base44 workflows runs --status failed` and `base44 workflows list`. SKILL.md says function logs are the wrong tool when a scheduled job did not run.
Try: “Why did my scheduled job fail last night?”
Runs code · Workflow check
Pulls the latest error-level lines across all functions
Command: `npx base44 logs --level error`.
Runs code · Recent errors
Filters logs to one function by name and level
SKILL.md warns that a --function filter can hide unstamped rows from apps on legacy deployment; re-run without it before concluding there are no logs.
Runs code · Function drill-down
Fetches logs between two moments with --since and --until
Both take ISO datetimes or shorthand such as 1h, 30m or 2d, per the reference.
Runs code · Time range
Treats "No logs found" as unknown and waits instead of changing flags
SKILL.md says an empty result usually means the run is not ingested yet, so change the time, not the flags.
Empty-result rules
Reads preview logs by default; --env prod reads the published app
If prod shows no logs the app may not be published yet.
Runs code · Environment choice

From the official docs

Permissions

Low risk

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

Logs end up in your Claude session
Function logs show what your functions printed. That text, including any user data your code logs, goes into the conversation. Filter with --function or --level when you can. github.com
It runs the base44 package through npx
Every command starts with npx base44. If you are not signed in, Claude stops and asks you to run npx base44 login yourself. github.com
It relies on base44-cli reference files
The workflow commands link to files inside the base44-cli skill. The base44-cli SKILL.md describes workflows list and workflows runs as read-only. github.com

Technical details

Turning on read-only	There is no switch. The skill only lists log and workflow commands, and the base44-cli SKILL.md describes the workflow ones as read-only.
Sign-in and keys	Uses the CLI login you already have, or --app-id and BASE44_APP_ID to pick the app.
Scopes	Runs npx base44 whoami, logs and workflows commands with your Base44 login Reads function logs and workflow runs for one app

Questions

What does the base44-troubleshooter skill do?

It makes Claude debug a Base44 app in production by reading function logs and workflow run history with the base44 command line. It covers watching a failure live, finding failed scheduled jobs and drilling into one function.

How do I install base44-troubleshooter?

Use the install box above. The plugin route installs it with base44-cli and 10 other skills. With the skills CLI, name base44-cli too: the skill links to workflow reference files that live in base44-cli.

Why does Claude say "No logs found"?

SKILL.md says that message is ambiguous and never means healthy. Most often the run is not ingested yet, which takes about 20 to 30 seconds, so wait and re-run. It can also mean the function name is wrong or the app has not been published when you read prod logs.

Should I use logs or workflow runs for a scheduled job?

Start with workflow runs. SKILL.md says function logs cannot show whether a run was dispatched at all or which task failed. Use base44 workflows runs --status failed first, then logs for the function a failing task called.

Do I need a Base44 login?

Yes. Claude checks with npx base44 whoami and asks you to run npx base44 login if you are signed out. You can also pass --app-id to read an app outside a linked project.

Is base44-troubleshooter safe?

It bundles no scripts and only lists log and workflow commands. Logs can contain whatever your functions printed, and that text enters your Claude session. Our advice: filter by function and level so less log text is pulled in.

Does it work in the Claude app?

The README documents Claude Code, Codex CLI, Antigravity and other agents through the skills CLI. It documents no Claude app upload route, and the skill needs a terminal for npx base44.

Is it an official Base44 skill?

It is in Base44's own repo, and the repo's marketplace.json describes the skills as "Official Base44 skills for Claude Code". The README marks them as beta.

Information

updated Nov 2, 2026

Repository: base44/skills
Folder: skills/base44-troubleshooter

base44/skills skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 6

base44/skills marketplace.json · github.com
base44 plugin.json · github.com
base44-troubleshooter SKILL.md · github.com
base44/skills README · github.com
base44 logs reference · github.com
base44-cli SKILL.md · github.com