Skip to main content

​
Fetch Logs

Retrieves raw logs from configured log sources without interpretation.

​
Overview

This skill fetches raw log data from discovered sources. It handles connection to various log storage systems and returns unprocessed log entries for further analysis.

​
When to Use

Use this skill when:
  • Raw logs are required for analysis
  • Time-bounded log retrieval is needed
  • Setting up the analysis pipeline
  • After log sources have been discovered

​
Directory Structure

fetch_logs/
├── SKILL.md
└── scripts/
    └── run.py

​
Instructions

  1. Receive log sources: Accept output from logsource_discovery
  2. Validate inputs: Ensure time range and source configs are complete
  3. Run fetcher: Execute the fetch script
  4. Outputs: Return raw logs in structured format
  5. Pass to parser: Provide raw logs to parse_logs
  6. Handle errors: Return empty result with error details

​
Input

{
  "sources": [
    {
      "type": "elasticsearch",
      "host": "es.example.com",
      "port": 9200,
      "index": "auth-service-*"
    }
  ],
  "time_range": "24h",
  "severity_filter": "ERROR",
  "max_logs": 10000
}

​
Output

{
  "logs": [
    {
      "timestamp": "2026-02-10T14:32:15Z",
      "raw_message": "ERROR: Database connection timeout",
      "source": "auth-service",
      "metadata": {
        "host": "prod-01",
        "file": "/var/log/app.log"
      }
    }
  ],
  "total_fetched": 1234,
  "time_range": "24h"
}

​
Example Request

uv run .agents/skills/fetch_logs/scripts/run.py --sources config.json --time-range 24h --max-logs 10000