fix(validation): retain the resolved Core Csc failure in history

[hcoona]

Summary

The failed Core Csc observer left action 0056 without a Windows completion receipt,
so the current history readers cannot admit later work. Record the independently
accepted current-bootstrap absence and retain this exact failure in the existing
handoff, preserving its original failure flags and consumed build/test unit.

Authorization and Governing Records

Issue #108 and the
accepted Windows Slice Wave entry
authorize credential-free implementation and validation preparation. This proposal
uses accepted target a42b6a648d4dfe4dc542256b9a8cc3faf70e0bd7, the experiment safety
policy, and the existing Windows Slice validation protocol. PRs #169 and #170 own
the preceding failure recovery and sole bootstrap-query procedure.

Scope and Non-Goals

Update the existing protocol and its three current Python consumers. Recognize
only disposed 0056 before ordinary success/inventory checks; select successful
guard 0055 explicitly; preserve historical reader pins using retained original
reader paths while keeping current-source admission separate. All new source
bindings remain unset pending exact admission.

The final caller's second failed-history checkpoint must also follow the new
successor number 0057, preserving its exact two-pass schedule.

The compiler graph, Native AOT artifact, final CLI/WSL scenarios, and real-account
acceptance remain incomplete. This change grants no observer/query retry or
reader execution and preserves all final publication gates.

Record-System Impact

The existing experiment protocol remains the canonical authority. Its existing
handoff gains the exact failed-observer suffix; private evidence instances bind
that authority and use the existing handoff acceptance carrier. No record family,
control, historical validator, or original receipt changes.

Evidence and Reasoning

The sole query returned one complete 160-byte line: absent, query completed, zero
rows/exact/ambiguous matches, and 1,630 ms internal elapsed time. Its original
invocation completed with exit zero in a conservative 2,687.223 ms interval. The
independent review accepts only current exact-bootstrap absence. Combined with
the accepted original-source reachability exclusion, it resolves the bounded
current lifetime question; it does not recover the historical exception or prove
observed historical no-launch.

The fixed failed-history pass reads the exact disposition and six original roles:
seven content reads, at most 15,093 requested bytes and 61 metadata calls. The
ordinary reader has one 30-second pass; final publication has two existing
checkpoints sharing 30 seconds of active verification, stable identities/bytes,
and the original outer deadline/cancellation. Failure latches. No complete 0056
inventory or missing Windows receipt is asserted.

Original history and guard evidence are preserved. The combined build/test debit
is 87/120, including the original fixture once. Ordinary Windows capacity stays
48/48; its aggregate 49 includes the dedicated observer unit. No capacity resets.

Identity and Security Effects

No account enumeration, token, consent, cache, broker, resource, telemetry, or host
state operation is introduced. The committed evidence contains sanitized outcomes
and hashes. Private evidence remains outside Git; original failure flags remain
unchanged.

Validation

The complete CI-equivalent hk check --all passed, including all 32 public-build
runner conformance groups, schema/catalog/identifier checks, links, whitespace,
secret scanning (current files and history), and actionlint.
Candidate consumer parsing, import, execution, and original-root reads have not
been performed; they retain their separate exact-admission requirements.

Review and Disposition

Independent query-outcome and prospective private-data reviews by wave_review
found no material findings. The collector's current-versus-historical reader-path
finding was independently triaged by wave_review as a true positive (confidence
10) and resolved by passing the retained original readers to the unchanged
historical validator. Final review by wave_review found a stale second-checkpoint
reservation expectation (0056) after the new history derives 0057. Independent
triage by lifetime_triage classified it as a true positive (confidence 10); its
one-literal correction is applied and the complete hk check passed again, including
all 32 conformance groups. Final independent source, research-evidence, and
record-system review by wave_review found no material findings on the corrected
snapshot. The sealed review is 8,123 bytes, SHA-256
6edfb64d0511c304bb0456ed88523a1de419ef5717ceed931ca200f6df81c7dc;
its independent finding triage is 4,159 bytes, SHA-256
c9b8e9e301c451c9aa7eb1faa8226a82ab8b266302a79a71cc423c0ea885a4ea.
The reviewer and author separately verified committed head
6908ab3b501a801f18016bb33f04f71b08d1e503, tree
a35ee4f45a440f09c7da0e70cb1b237b4d885d71, against that exact reviewed snapshot.
No owner decision is pending and no recheck trigger or Wave change is introduced.

Upstream Provenance

No new upstream source is copied or derived. Existing compiler and framework
provenance remains unchanged.