Data
20Slice, filter, transform, and query structured data: JSON, YAML, XML, CSV, and SQL. The glue that turns raw output into something the agent can reason over.
JSON
- jqsed/awk for JSON; filter, transform, and query JSON streams. The canonical agent JSON tool.
- gojqpure-Go jq with better error messages; drop-in replacement.
- fxJSON viewer and processor with JS expressions; scriptable transforms.
- jlessread-only viewer for exploring and searching large JSON payloads.
- gronflatten JSON into greppable
path = valueassignments and back; makes JSON searchable with grep. - jcconvert the output of common CLIs (ls, ps, dig, ifconfig...) into JSON. Turns arbitrary command output into structured data an agent can parse.
YAML / XML / multi-format
CSV / tabular
- Millerawk/sed/cut/join/sort for CSV, TSV, JSON, and JSON Lines; streaming tabular transforms.
- csvkitsuite to convert, query, and manipulate CSV files.
- qsvextremely fast CSV query/slice/join/validate/transform toolkit for large data.
- xsvfast CSV indexing, slicing, joining, and stats.
- VisiDatadata multitool with a batch/replay mode for scripted tabular processing.
SQL over files & embedded databases
- DuckDBin-process analytical SQL over CSV/Parquet/JSON and databases; blazing fast, zero setup. A default for agent data work.
- SQLitethe embedded SQL database;
sqlite3runs queries against local.dbfiles non-interactively. - sqlite-utilsbuild, query, and manipulate SQLite databases from the CLI; great for agents assembling structured stores.
- sqjq-style access across SQL databases, CSV, and Excel; unified querying of structured sources.
- qrun SQL directly against CSV/TSV files as if they were tables.