git extdeps is wrong - lets find the scope

[gunbai-bot]

Auto-opened by session-dashboard for session zesty-lark-54.
Pushing to session/zesty-lark-54 advances this PR.

Worker attestation

Before flipping this PR to ready for review, confirm each item:

• Title describes the change (not the session id or branch).
• PR body summarises what and why (replace the TODO below).
• Tests run: name the command (e.g. npm test, cargo test) and the result.
• If this closes a work item, the body contains a Closes #N directive.
• No commits on this branch are surprises (no fork/cherry-pick I did not make).
• No secrets / credentials / large binaries staged.

Summary

TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.

Test plan

• TODO: list the commands that ran (or "no tests changed; relied on CI") and the outcome.

[gunbai-bot]

CI triage — the prior red was a main-skew, not this PR's change; this PR now carries the repair.

The failing check (claim_batch floor) was name 'grammar_relation_field_tokens' not found in module 'v2.std.compilers.target_model', from a #5077 × #5082 semantic merge-skew on main (no merge queue, two individually-green PRs combine RED):

• Ban self-naming symbol bindings: repo-wide sweep (data X: Symbol = X → ^X) #5077 ("Ban self-naming symbol bindings: data X: Symbol = X → ^X", repo-wide sweep) deleted the named binding grammar_relation_field_tokens from target_model.
• Realize the grammar-inverse (.dag unparser) for ONE shape — a type-decl Conj node — proving parse(emit(n)) ≡ n by execution. This is NOT a new unparser: emission = ingestion⁻¹ over the SAME GrammarRelation authority (a row carries grammar_relation_field_tokens AND grammar_relation_field_emitted; 03_ #5082 (grammar-inverse type-decl round-trip test) merged after Ban self-naming symbol bindings: repo-wide sweep (data X: Symbol = X → ^X) #5077, so its new emit_ingest_type_decl_round_trip_test.dag was never swept — it still imported the deleted binding. → main HEAD (1bffd575) is red; every PR's merge-commit CI inherits it.

This PR's merge of main applies #5077's exact pattern to that one straggler file (4-line change, no substrate/proof-logic change):

• import { … grammar_relation_field_tokens } → use the ^grammar_relation_field_tokens symbol literal at the find_named_child call site;
• removed the one self-naming binding Realize the grammar-inverse (.dag unparser) for ONE shape — a type-decl Conj node — proving parse(emit(n)) ≡ n by execution. This is NOT a new unparser: emission = ingestion⁻¹ over the SAME GrammarRelation authority (a row carries grammar_relation_field_tokens AND grammar_relation_field_emitted; 03_ #5082 re-introduced (the file Ban self-naming symbol bindings: repo-wide sweep (data X: Symbol = X → ^X) #5077 had driven to zero) — its ^type_decl_round_trip_foreign_binding literal was already used at the call site.

Verified green by execution (not grep): the previously-red floor step now reports PASS emit_ingest_type_decl_round_trip_holds (resolves clean, 61 modules) and PASS corpus_gcp_oauth_defused_holds (this PR's de-fusion guard), v2 claim corpus: 198 discriminating witnesses passed, exit 0.

Heads-up on merge ordering: main is red on this skew right now. Merging this PR repairs main; alternatively the same 4-line straggler fix can be applied directly to main.