# Isolated default-policy experiment

Read [the implementation/evidence report](../../text-default-omission-policy-20260913.md).
The original iteration passed16 map tests and strict targeted Clippy. After
review, two tests were added; the final iteration passes18 and strict targeted
Clippy. Prototype/production bytes stayed unchanged between iterations.

Each of the four gate directories contains raw stdout/stderr, start/result
receipts and608 captured source files in JSON/ZIP. The capture explicitly
includes the compiled harness prototype. The exact runner is `gate-policy.py`;
its producer-host paths need adaptation for a new capture. The report supplies
direct reproducible Cargo commands. Reviews retain their static/runtime scope.

Run `python -B verify.py` here for read-only checks of membership, byte hashes,
source ZIPs, receipts, Cargo outcomes, the two added test names and unchanged
prototype/production across iterations. `files.json` pins every other file;
Git identifies that manifest. Test binaries are not included; hashes are
historical runner receipts. These records establish an experimental map policy,
not public source formatting, lossless edits, source8, GUI/API, Linux or release
acceptance. The production baseline's three preservation defects remain open.