LoggedWebSearch

DuckDuckGo Search (Logged)

Search the web using DuckDuckGo and automatically save results to a plain-text log file.

When to use

Use this skill when the LoggedWebSearch name is included in the user query.

Execution

This skill runs a Python script and parses its JSON output.

Script location: workspace/Skills/LoggedWebSearch/logged_web_search.py

Command pattern:

python logged_web_search.py --query "<search_query>" --max <number> --timeout <seconds> --sleep-ms <milliseconds>

Command-Line Arguments

All arguments except --query are optional.

• --query (required, string)

  • The search query to send to DuckDuckGo
  • Example: "python programming"

• --max (optional, integer, default: 8)

  • Maximum number of results to return
  • Valid range: 1 to 25
  • Values outside range are clamped

• --timeout (optional, integer, default: 20)

  • Network timeout in seconds
  • Valid range: 5 to 60 seconds
  • Values outside range are clamped

• --sleep-ms (optional, integer, default: 0)

  • Milliseconds to sleep before searching
  • Useful for rate limiting
  • Example: --sleep-ms 500 waits 0.5 seconds

Response Format

The script outputs valid JSON to stdout.

Success Response (status: "ok")

{
  "query": "search query string",
  "results": [
    {
      "title": "Result Title",
      "url": "https://example.com",
      "snippet": "Brief description of result",
      "rank": 1
    }
  ],
  "log_path": "/path/to/logfile.txt",
  "count": 5,
  "status": "ok"
}

Error Response (status: "error")

{
  "query": "search query string",
  "error": "Error message describing what went wrong",
  "results": [],
  "count": 0,
  "status": "error"
}

Logging

Search results are automatically saved to plain-text files at:

{OPENCLAW_OUTPUT_ROOT}/datastore/YYYY/MM/DD/{QueryName}.txt

Where:

• YYYY/MM/DD is the current date (e.g., 2026/02/17)
• {QueryName} is the query converted to CamelCase (e.g., "PythonProgramming")

If OPENCLAW_OUTPUT_ROOT environment variable is not set, defaults to:

~/.openclaw/data

Each log entry includes:

• Query text
• ISO timestamp
• Result title, URL, and snippet for each result

Examples

Example 1: Simple search

python logged_web_search.py --query "climate change"

Returns up to 8 results (default).

Example 2: Limited results with timeout

python logged_web_search.py --query "python asyncio" --max 5 --timeout 15

Returns up to 5 results with 15-second timeout.

Example 3: Rate-limited search

python logged_web_search.py --query "AI news" --sleep-ms 1000

Waits 1 second before searching.

Troubleshooting

• No output from python logged_web_search.py

  • Confirm workspace/Skills/LoggedWebSearch/logged_web_search.py exists.
  • Verify Python is available in the sandbox (python --version or python3 --version).

• No log files created

  • Ensure OPENCLAW_OUTPUT_ROOT points to a valid, writable path.
  • If unset, the default ~/.openclaw/data must be writable and exist.

• Cannot list or create files

  • The execution environment may be sandboxed; only declared filesystem capabilities are available.

• Network timeout: Increase --timeout value

• No results: Try a simpler query with fewer words

• Connection refused: Check internet connectivity

• Log file not created: Verify OPENCLAW_OUTPUT_ROOT environment variable or .openclaw/data directory exists