HOW-TO-RUN.md

The command reference for the analyst's tools.

What it holds

Example command lines in groups: quotes and the store (cite.py and the mem.py verbs, those absent in v37 marked new); the fleet's pages (packets.py); the return (handoff.py, ingest.py, rebuild.py); open, close, cut and views (validate.py, gen_views.py, done.sh, snap.py, counts.py). Then "The harness, as it concerns you": arm.json, legs, the head at most 3,000 words and "rewritten at every close and before every cut", compaction, three limits of the runner, forks. Last, stage 00's builders. It says every mem.py call is logged and pages are 80 lines.

Who writes it, and when

Written with the instrument; build.sh copies it to the analyst's root. It is in A1's first-leg packet. STANDING §1 and A1 name it tools/HOW-TO-RUN.md; build.sh places it at the root and leaves any copy in tools/ out.

Why it exists

STANDING §1 calls it "the reference for all of them" and says to read it once at A1. Where it and STANDING's list of verbs added in the last tools pass differ, that list wins.

Where

HOW-TO-RUN.md; source aggregator/instrument/HOW-TO-RUN.md.

Checked against: HOW-TO-RUN.md, STANDING-INSTRUCTIONS.md §1, stages/STAGE-A1-reading-room-notes.md, harness/build.sh.