This directory contains research utilities, runners, diagnostics, and proof-board builders. It is intentionally broad, but the public review path is narrow.
| File | Why it matters |
|---|---|
test_live_kraken_spot_frontier_maker_machinegun_shadow.py |
Focused unit suite used by GitHub Actions. |
live_kraken_spot_frontier_maker_machinegun_shadow.py |
Runner covered by the focused test suite. |
kraken_spot_client.py |
Exchange client wrapper used by Kraken spot workflows. |
live_penetration_lattice_shadow.py |
Shared shadow-runner support surface. |
process_singleton.py |
Runtime guard for avoiding duplicate runner processes. |
benchmarks/ |
Benchmark harnesses moved out of the public repo root. |
- Treat
test_*files as behavior specifications or regression probes. - Treat
build_*,watch_*, andvalidate_*files as evidence and operator-surface generators. - Treat live or broker-connected scripts as local-only workflows unless explicitly run in dry-run or shadow mode.
- Do not expect every historical research script to be a polished command-line product.
For the quickest technical review, start with the focused test and runner pair above, then compare their behavior to the proof snapshot in ../docs/evidence/edge_registry.md.