From repeated friction
to a governing rule.

The review refusal states the obligation before a formal protocol exists: add the missing tests, provide the relevant test file and evidence, and request review.

These events were visible to the proposer. Their order is recorded; their exact steps are not quoted in the manuscript.

P01 applies to a PR affecting public callables or cross-module behavior. Its review record must identify every changed callable signature and every added or modified cross-module invocation.

Each item must map to the applicable written contract. Required public smoke tests must run through the highest stacking step affected. Unmappable items and required failures must be resolved or explicitly rejected before merge approval.

The ledger records support and adoption at step 21. The compiled specification is registered in protocol state and supplied to the action context.

Adoption is a separate event from proposal and from subsequent use. Other proposals in the same run were not adopted.

A declared-action use does not by itself establish an independently verified change to a decision or execution path. The manuscript records use and enforcement separately.

The protocol-linked refusal is retained with paired snapshots of the governed object.

PR-A is followed below to step 138. Its eventual resolution requires intervening implementation and verification; the trace does not isolate how much delay the rule caused.

The recorded enforcement rate changes from 0.50 to 0.68 events per step across this amendment. These rates describe the rule’s own ledger activity, not an independent outcome.

The manuscript preserves the amendment source AM-A and its two edits. These are amendment-draft events, distinct from the final change to the active protocol.

The recorded enforcement rate changes from 0.55 to 0.69 per step across the second amendment. This is a descriptive ledger comparison, not a causal outcome estimate.

If mainline changes after either result is recorded, synchronize the PR and rerun both checks on the resulting current head.

A reviewer or merger must be able to compare the recorded tested commit with the current head and determine pass/fail without relying on the author’s statement. P04 records 247 uses and four amendments by the endpoint.

P05 is a readiness rule, not an identical copy of P04’s head-binding specification. It records 242 uses and seven amendments by the endpoint.

That observed interval includes additional implementation, verification, and coordination. It is not an estimate of extra delay caused by the rule.

Across the twelve affected PRs, eleven later merge, with first-block-to-merge intervals of 15–112 steps.

P06 records 39 uses and zero reported enforcements. Its positive use count is distinct from the principal rule’s enforcement record.

P09 has only 45 steps of post-adoption observation before step 336 and cannot meet the paper’s 48-step persistence threshold within this run.

Its inventory shows no numeric use, enforcement, or amendment count. A missing count is not treated as zero or as evidence of sustained use.

The source reports thirteen amendments across P01, P04, and P05, with the last at step 317. It does not identify that final amendment’s protocol in the displayed timeline.

Of 76 total merges, 65 were never blocked by P01. The selected trace does not identify how much of the run-level delivery difference this rule caused.

The B3 endpoint passes 28 of 35 behavioral cases and completes 2 of 5 contracts. These are whole-run outcomes, distinct from the single-rule trace.

The case has no human participant. PR-B has no observed override at the horizon.