docs(remediation): re-run failed; RM-02 blocked on RM-61, declared bound honoured

Pipeline #2188 at the same head: ci-postgres the sole failure, everything else success
including gate-verify. f10-coder independently confirmed and applied the pre-registered
rule itself.

RM-02 blocked. No ad-hoc exemption, no third roll. Recorded why: two failures against one
earlier clean run makes 'best of five' feel reasonable, and the point of declaring the
bound in advance is that it binds when the result is inconvenient. A third roll would
have been indistinguishable from diligence, including to the person doing it.

Data for RM-61: the artifact hit twice consecutively at f9746b23 while #2184 was clean at
38f1b249, so it may cluster rather than be uniformly random — which makes a one-shot
re-roll close to worthless and strengthens the ruling that it must not become policy.

And the illustration: the anti-inert-gate registry is blocked by an artifact that makes
gate evidence unreliable. The thesis demonstrating itself on its own keystone.

Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>

docs/remediation/TASKS.md

@@ -249,6 +249,36 @@ discrimination, not a retry mechanism.
 **A clean re-run does NOT retire RM-61.** The re-roll is a die whose bias remains unmeasured; only RM-61's
 negative control measures it. RM-61 stays required regardless of this run's outcome.
 
+> **OUTCOME — the re-run FAILED, and the declared response was honoured.** Pipeline **#2188**, same head
+> `f9746b23`, JSON scan 11 entries: `ci-postgres` the **sole** failure, every other child success
+> including `gate-verify`. Confirmed independently by `f10-coder`, which applied the pre-registered rule
+> itself — _"I do not call terminal-green, do not request a third roll, and do not infer an exemption."_
+>
+> **RM-02 is BLOCKED on RM-61.** No ad-hoc exemption. **No third roll.**
+>
+> Worth stating why not, because the temptation was legible: two failures against one earlier success
+> (#2184 was clean at the previous head) makes _"best of five"_ feel reasonable. **The entire point of
+> declaring the bound in advance is that it binds when the result is inconvenient.** A bound honoured
+> only when it agrees with you is not a bound — and a third roll would have been indistinguishable from
+> diligence, _including to the person doing it_.
+>
+> **Data point for RM-61:** the artifact hit **twice consecutively** at `f9746b23` while #2184 was clean
+> at `38f1b249` — so it may **cluster** rather than being uniformly random per run. That makes the single
+> bounded re-roll a _weaker_ mitigation than it appeared, and strengthens the ruling that it must never
+> become policy: **if failures cluster, a one-shot re-roll is close to worthless.**
+
+### The sharpest illustration the mission has produced
+
+**RM-02 — the anti-inert-gate registry, built so a gate that has silently stopped enforcing can be
+DETECTED — is itself blocked by an infrastructure artifact that makes gate evidence unreliable.**
+
+Its code is complete. Its own harness passes in CI. Two independent reviews cleared its substance. It
+cannot be certified terminal-green because the CI record contains a failure **nobody can yet prove is
+meaningless**.
+
+That is not an obstacle to the mission. It is the mission's thesis demonstrating itself on its own
+keystone deliverable.
+
 ### ★ MILESTONE — the full merge stack completed end-to-end for the first time (RM-03 / PR #1032)
 
 `independent review → CI terminal-green at the exact head → merge-gate verdict → coordinator → held for owner`