XL-2: record the caret handoff to the MQ-1 symbol-literal slice on its RFM row

[gunbai-bot]

Carrier-only change, ruled by quiet-seal-543 (option B: park the caret lowered form in XL-2 and hand it to the MQ-1 symbol-literal slice).

It adds one receipt to gunbc.recurring_failure_mode caret_symbol_has_no_lowered_form that records:

• The handoff, and why a located refusal is enough for XL-2.
• The three findings that bound the climb: the name text is absent from the tree by design, the target is std.syntax LitSymbol, and Symbol typing is step (a) of symbol_identity_census.
• A correction the row lacked: the in-form let-value instance, measured on main 88436e5, is accepted with the caret dropped. It is owned by first-match PR-2 (bright-boar-848).

No code changes. The row's rung is unchanged.

🤖 Generated with Claude Code

[gunbai-bot]

CI red at f67487e is not this change, which edits only an RFM row string. The only blocker is the wet-lane witness test.claim.approval_ntfy_access_readback_wet_witness_test.the_root_refuses_before_minting_on_this_host_by_real_execution (expected passed, observed failed). It executes on the runner host. The same identity failed on an unrelated PR (run 36562259065, fierce-deer-555) and passed on merge-group run 36562310194. It is therefore intermittent across runners, which does not by itself prove it environmental. I reran the failed jobs. — sent from sharp-carp-336

[briansrls]

APPROVE-MERGE at exact head f67487e, through the merge queue only.

This is carrier-only and changes one receipt on caret_symbol_has_no_lowered_form. The handoff is bounded correctly: XL-2 keeps the located-refusal floor, while the actual lowered form belongs to MQ-1's symbol-literal slice. The receipt identifies the existing target authority (std.syntax LitSymbol), records that caret identifiers are currently class-stamped rather than lexeme-stamped and therefore require the span/lexeme step first, and binds Symbol typing to symbol_identity_census step (a) rather than opening a second decision or literal authority.

The correction to the historical receipt is also appropriate: the in-form let value was a separate first-match silent-drop route, and it is attributed to #12615's let-in value reader rather than claimed as repaired by this carrier PR. No rung or implementation claim changes here.

Exact-head witnesses run 36563055996 completed successfully, GitHub reports the PR mergeable, and the diff is exactly one added receipt in the RFM row. No local tests were run by me. Require the actual merge_group candidate to pass against then-current main; no direct merge or check bypass.