tri rule-observance: measure the practice, not the seal (Refs #2622)

[gHashTag]

tri loop-rules checks docs/loop/LOOP-RULES.md against a sealed digest and passes. It closes with its own disclaimer:

"This certifies identity only. It does not certify that the rules are correct, nor that the tick obeyed them."

That is the file's own R6 — the seal certifies identity, not correctness — stated about the tool that enforces it. Nothing measured the second half.

Over the last 30 merged pull requests, every author [measured]:

clause (R11)	observed
branch prefix w699-<topic>	0 of 30
provenance tag on a number	0 of 25 stating one
no Closes #N autoclose	30 of 30 ✓

The tag convention is alive elsewhere — [measured] appears 211 times in the theory document. It has never reached a pull request.

Two clauses at zero across every author is not one operator drifting. The file and the practice have parted, and the seal made the file look governed while nothing looked at the practice — a checksum over rules is exactly as strong as a green check over a gate that cannot fail.

Reports, does not gate

Two separable reasons:

• Which way a drift resolves is not the measurer's call — practice to rule, rule to practice, or rule plus a gate. Encoding a preference decides it by default.
• Most of the file is not mechanically checkable. "Your own instrument is the first suspect" is a judgement about a tick's content; a regex pretending to score it would make the unmeasured look measured, which is this campaign's whole subject.

Filed as #2622, including the clause I knowingly do not observe — "merging is a human act" — which the owner authorised outside this file. The rule is not edited to match what I do: a rules file edited by the thing it governs is worth nothing.

And the smaller finding under it: R2 is "your own instrument is the first suspect." I have re-derived that from scratch five times this campaign while it sat written down, sealed, in a file I had never opened.

§104.

Refs #2622

[github-actions]

📓 NotebookLM Notebook linked to this PR

• Notebook ID: b83263109fb055dc
• URL: https://notebooklm.google.com/notebook/b83263109fb055dc

This notebook contains session context, decisions, and artifacts for this work.

[github-actions]

PR Dashboard

Generated at: 2026-08-23 22:40:16 UTC

• tri rule-observance: measure the practice, not the seal (Refs #2622) #2624 tri rule-observance: measure the practice, not the seal (Refs #2622) | Created: 2026-08-23 | Updated: 2026-08-23 | 18 failing / 22 total
• Unbreak FPGA CI and correct two unbacked hardware claims #2081 Unbreak FPGA CI and correct two unbacked hardware claims | Created: 2026-08-11 | Updated: 2026-08-20 | 5 failing / 27 total

Summary

Status	Count
Total Open PRs	2
PRs with Failing Checks	1
PRs with All Checks Green	1
READY	0
FAILING	1
PENDING	0

Seal Status

• ⚠️ STALE -- sha256(compiler.rs)=e6333575ec10 != manifest seal=87e5cbd3ad94.
  The committed NMSE numbers were certified against an older compiler.rs.
  Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).