Slash Command

/llm-externalizer-mass-scout-export

Dump every result row of a mass-scouting job to JSONL or CSV under reports/mass_scouting/. Useful for follow-up analysis in pandas, jq, etc.

npx claudepluginhub emasoft/emasoft-plugins --plugin llm-externalizer

Details

Argument--db <path> --job-id <id> [--format jsonl|csv]

Allowed Tools

mcp__llm-externalizer__mass_scout_export

Prompt Preview

# Mass-scout — export

Iterate every `mass_scout_results` row for a given `--job-id` and write
them to a single JSONL or CSV file under
`<main-repo-root>/reports/mass_scouting/`. The file path is returned to
the caller so they can route it to the next stage of their workflow.

## Inputs

| Flag | Required | Description |
|---|---|---|
| `--db <path>` | yes | The registry |
| `--job-id <id>` | yes | The job to export |
| `--format jsonl\|csv` | no | Default: `jsonl` |

## Output



## JSONL format

One JSON object per line — same shape as the SQLite `mass_scout_results`
row:



## CSV format...

Command Content

Stats

Stars1

Forks0

Last CommitMay 7, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

# Mass-scout — export Iterate every `mass_scout_results` row for a given `--job-id` and write them to a single JSONL or CSV file under `<main-repo-root>/reports/mass_scouting/`. The file path is returned to the caller so they can route it to the next stage of their workflow. ## Inputs | Flag | Required | Description | |---|---|---| | `--db <path>` | yes | The registry | | `--job-id <id>` | yes | The job to export | | `--format jsonl\|csv` | no | Default: `jsonl` | ## Output ## JSONL format One JSON object per line — same shape as the SQLite `mass_scout_results` row: ## CSV format...

Mass-scout — export

Iterate every mass_scout_results row for a given --job-id and write them to a single JSONL or CSV file under <main-repo-root>/reports/mass_scouting/. The file path is returned to the caller so they can route it to the next stage of their workflow.

Inputs

Flag

Required

Description

--db <path>

yes

The registry

--job-id <id>

yes

The job to export

--format jsonl|csv

Default: jsonl

Output

job_id=<id> format=jsonl|csv rows=<N> path=<absolute path to the export file>

JSONL format

One JSON object per line — same shape as the SQLite mass_scout_results row:

{"job_id":"...","file_fingerprint":"...","short_id":1,"result_json":"...","raw_response":"...","repaired":0,"attempts":1,"cost_usd":0.000_005,"enriched_at":"2026-05-06T..."}

CSV format

Columns: job_id, file_fingerprint, short_id, result_json, repaired, attempts, cost_usd, enriched_at

result_json is the embedded JSON (already a string). Cells that contain commas, quotes, or newlines are quoted and double-escaped.

Tips

jq -c '.result_json | fromjson | .is_async' <file>.jsonl — extract one field across all rows.

Pipe CSV into pandas: pd.read_csv(...) then pd.json_normalize(..., record_path=None, meta=...) on result_json to expand the dynamic fieldset into proper columns.