name	logbasset
description	Query, search, aggregate, or live-tail Scalyr/DataSet logs from the command line with the logbasset CLI. Use when the user wants to investigate logs, find events by ID or field, count or aggregate log data over a time range, or live-tail a log.

logbasset

logbasset is a read-only command-line tool for querying Scalyr/DataSet logs. It can search raw log records, run aggregations (counts, groupings, facets, time series), and live-tail a log. Nothing it does mutates data.

This skill does not restate the command and flag reference, because the binary documents itself and stays current with the installed version. Load that reference at runtime instead of guessing.

Workflow

Follow these steps in order:

Check the binary is installed. Confirm logbasset is on PATH (e.g. logbasset --version). If it is not found, stop and tell the user to install it (brew install andreagrandi/tap/logbasset, or download from https://github.com/andreagrandi/logbasset/releases) — do not attempt to query without it.
Load the authoritative reference. Run logbasset context to print the current list of commands, flags, time formats, output formats, exit codes, and worked workflows for the installed version. Treat this output as the source of truth.
Confirm exact flags before composing a non-trivial query. Run logbasset schema <command> for a specific command's flags, enums, and defaults, or logbasset schema global for the flags shared by every command. Do not invent flags — logbasset context lists ones that look plausible but do not exist.
Construct and run the query using only flags confirmed from the steps above.

Safety

Every logbasset command is read-only — queries never create, modify, or delete data — so they are safe to run without asking the user for confirmation.

Keeping queries fast and inexpensive

Prefer narrow time ranges (--start 1h over --start 7d).
Cap raw-record fetches with --count.
Prefer aggregation commands over fetching many raw records when the user wants counts, rates, or top values.
Use --priority low for heavy or background queries so interactive ones stay responsive.

For the exact flag names, output formats, and time syntax, always rely on logbasset context and logbasset schema rather than memory.

logbasset

This skill does not restate the command and flag reference, because the binary documents itself and stays current with the installed version. Load that reference at runtime instead of guessing.

Workflow

Follow these steps in order:

Check the binary is installed. Confirm logbasset is on PATH (e.g. logbasset --version). If it is not found, stop and tell the user to install it (brew install andreagrandi/tap/logbasset, or download from https://github.com/andreagrandi/logbasset/releases) — do not attempt to query without it.

Load the authoritative reference. Run logbasset context to print the current list of commands, flags, time formats, output formats, exit codes, and worked workflows for the installed version. Treat this output as the source of truth.

Confirm exact flags before composing a non-trivial query. Run logbasset schema <command> for a specific command's flags, enums, and defaults, or logbasset schema global for the flags shared by every command. Do not invent flags — logbasset context lists ones that look plausible but do not exist.

Construct and run the query using only flags confirmed from the steps above.

Safety

Every logbasset command is read-only — queries never create, modify, or delete data — so they are safe to run without asking the user for confirmation.

Keeping queries fast and inexpensive

Prefer narrow time ranges (--start 1h over --start 7d).

Cap raw-record fetches with --count.

Prefer aggregation commands over fetching many raw records when the user wants counts, rates, or top values.

Use --priority low for heavy or background queries so interactive ones stay responsive.

For the exact flag names, output formats, and time syntax, always rely on logbasset context and logbasset schema rather than memory.