AI Coding Agent Skill

Name: uLogger Data Client
Author: uLogger

Give your AI coding agent direct access to the uLogger platform. Query logs, push device configurations, create firmware deployments, and manage certificates — all through natural language.

What the Skill Does

The uLogger Data Client skill teaches your AI agent how to use the ulogger-data-client CLI to interact with the uLogger platform on your behalf.

🔍 Log Queries

Search and filter device logs by time window, application, device serial, and log level. Triage incidents in seconds.

⚙️ Log Configuration

Push log level and module configurations to devices over MQTT. Enable specific debug modules by name.

🐛 Fix Field Issues

Query for hard faults, assert failures, and critical error logs across your fleet. Your agent identifies patterns, correlates them with firmware versions, and suggests or applies code fixes directly in your repository.

🔌 Bootstrap uLogger

Download your device certificate bundle, embed the keys as a C header file, and integrate the uLogger library into your embedded project — the agent follows the integration guide and wires everything up for you.

Download

Download the CLI binary for your platform. The same binary works as both a standalone CLI tool and as the executable backing the AI skill.

SKILL.md file: The skill definition file is also available for download: Download SKILL.md. Place it alongside the binary in your skill directory.

Installation

1 Download the binary

Download the correct binary for your operating system and architecture from the Downloads section above.

2 Download the SKILL.md

Download the SKILL.md file. This file teaches your AI agent how to use the CLI — what flags to pass, what workflows to follow, and how to interpret the output.

3 Create the skill directory

Create a directory for the skill and place both files inside it:

Windows (PowerShell)

mkdir "$env:USERPROFILE\.agents\skills\ulogger-data-client"
Copy-Item ulogger-data-client-windows-amd64.exe "$env:USERPROFILE\.agents\skills\ulogger-data-client\ulogger-data-client.exe"
Copy-Item SKILL.md "$env:USERPROFILE\.agents\skills\ulogger-data-client\SKILL.md"

macOS / Linux

mkdir -p ~/.agents/skills/ulogger-data-client
cp ulogger-data-client-<platform> ~/.agents/skills/ulogger-data-client/ulogger-data-client
chmod +x ~/.agents/skills/ulogger-data-client/ulogger-data-client
cp SKILL.md ~/.agents/skills/ulogger-data-client/SKILL.md

4 Set your API token

Set the ULOGGER_API_TOKEN environment variable so the CLI can authenticate without passing --token on every command. You can generate an API token from your uLogger dashboard.

Windows (PowerShell)

$env:ULOGGER_API_TOKEN = "ulg_YOUR_TOKEN_HERE"

macOS / Linux

export ULOGGER_API_TOKEN="ulg_YOUR_TOKEN_HERE"

5 Configure your AI platform

Point your AI coding agent to the skill directory. See the Configuration section below for platform-specific instructions.

AI Platform Configuration

Configure your AI coding agent to discover and use the uLogger skill.

VS Code + GitHub Copilot

Place the skill in the global agents directory. VS Code will automatically discover it across all workspaces.

# Global skill directory (all workspaces)
%USERPROFILE%\.agents\skills\ulogger-data-client\
├── SKILL.md
└── ulogger-data-client.exe

# Or per-project (workspace-specific)
.github/skills/ulogger-data-client/
├── SKILL.md
└── ulogger-data-client.exe

The SKILL.md frontmatter includes a name and description that Copilot uses to match your requests to the skill.

Cursor

Add the skill to Cursor’s rules directory so it can reference the SKILL.md when processing your requests.

# Add to .cursor/rules/ in your project
.cursor/rules/ulogger-data-client.md

# Contents: copy the SKILL.md content, or
# reference the executable path:
Executable: ~/.agents/skills/ulogger-data-client/ulogger-data-client

Point the executable path in the rule file to wherever you placed the binary.

Claude Code

Add the skill to your CLAUDE.md or .claude/skills/ directory so Claude can discover and use it.

# Place in your project or home directory
.claude/skills/ulogger-data-client/
├── SKILL.md
└── ulogger-data-client

# Or reference in CLAUDE.md:
See .claude/skills/ulogger-data-client/SKILL.md
for the uLogger Data Client CLI skill.

Other AI Agents

Any AI agent that supports skill files or tool definitions can use the uLogger Data Client. Provide the agent with:

# 1. The SKILL.md file (skill instructions)
# 2. The binary path
# 3. The ULOGGER_API_TOKEN env var

# The agent reads SKILL.md to learn
# the CLI flags, workflows, and output
# format — then runs the binary directly.

Usage Examples

Once installed, just describe what you want in natural language. Your AI agent translates requests into the correct CLI commands.

🔍 Query Recent Logs

Ask your agent:

"Show me the last hour of logs for device 1001"

Example output:

{
  "count": 5,
  "logs": [
    {
      "_msg": "Super-dee-duper bad error",
      "_time": "2026-05-21T01:40:39.44167Z",
      "application_id": "899391",
      "device_serial": "1001",
      "device_type": "WiFi_Demo",
      "level": "3",
      "version": "v2.0.0"
    },
    {
      "_msg": "Let's create an error-level log and observe this",
      "_time": "2026-05-21T01:40:39.434437Z",
      "device_serial": "1001",
      "level": "0",
      "version": "v2.0.0"
    }
  ]
}

→ 5 log entries returned

⚙️ Push a Log Configuration

Ask your agent:

"Enable debug logging for the SYSTEM and SENSOR modules on device 1001"

Example output:

{
  "success": true,
  "log_level": "DEBUG",
  "module_flags": 10,
  "timeout_seconds": 300,
  "topic": "config/v0/975773647/899391/1001"
}

📦 List Available Debug Modules

Ask your agent:

"What debug modules are available on device 1001?"

Example output:

→ 5 module(s) available:
  COMM             (bit 4)
  MAIN             (bit 1)
  POWER            (bit 16)
  SENSOR           (bit 8)
  SYSTEM           (bit 2)

Use --modules NAME1,NAME2 with --send-log-config to enable specific modules.

🔌 Integrate uLogger Into My Project

Ask your agent:

"Integrate uLogger into my embedded project"

Your agent downloads your device certificates and private keys from the uLogger platform, embeds them as C header files into your project, and follows the integration guide to bootstrap the uLogger library into your codebase.

Example output:

Downloading certificate bundle ...
Extracting to ./certs ...
  ✔ ulogger_certs_keys.h
  ✔ endpoint.txt
  ✔ thing_name.txt
  ✔ readme.txt
✔ Extracted 4 file(s) to ./certs
── Embedded / C/C++ Integration ──────────────
1. Copy ./certs/ulogger_certs_keys.h into your project's source tree.
2. Include the header in your TLS setup code:
     #include "ulogger_certs_keys.h"
   Available symbols:
     ulogger_user_certificate   — device certificate (PEM, null-terminated)
     ulogger_user_private_key   — device private key (PEM, null-terminated)
     ulogger_CA_certificate     — CA certificate (PEM, null-terminated)
3. Connect to the MQTT broker at mqtt.ulogger.ai:8883 using mutual TLS.

Ready to Get Started?

Create a free account to get your API token, then install the skill and start talking to your devices.

Create Free Account View Documentation