Document dissolution trigger for #6944's restored real_ingest_test imports

[briansrls]

Summary

• Main-red was misdiagnosed in the originating brief as sccache serving a zero-byte cached probe_bin. That log line is a passing negative-control test (witness_zero_byte_artifact_rejected_by_execution in dag/tools/build_step_transport.dag). The real failure, ~17 minutes later in the run log, was a whole-tree typecheck break: namespace wave 1: containment-tree resolution — layered census + name-derived loader (salvage) #6848 landed src/v2/test/claim/program_assembly/real_ingest_test.dag referencing several symbols without importing them, cascading into the batch-4 gate source_root_ingest_gate_passes.
• That import fix has since landed independently via fix(test): restore imports stripped from real_ingest_test.dag by #6848 #6944 (which restores the same imports, plus v2.std.diagnostic/v2.std.logic). This PR no longer carries the import diff — it is reduced to just the one thing fix(test): restore imports stripped from real_ingest_test.dag by #6848 #6944 didn't add: a DESIGN §6 dissolution-trigger note.
• The restored imports are a scaffold, not a permanent fix: v1's dependency-pool pull (resolve_transitively_with_seen in src/v1/stage0/src/main.rs) walks import-line text to build the compiled set, so a same-tree provider with no import edge silently drops out of the corpus even when it's otherwise reachable. That's the actual defect class namespace wave 1: containment-tree resolution — layered census + name-derived loader (salvage) #6848 tripped over. The added data real_ingest_test_import_restoration_note names the mechanism plainly, says the imports are temporary, and gives the dissolve-at condition (pool-pull becomes reference-driven / consumer-reachability-based instead of import-line-walk-based) — so the imports don't sit there indistinguishable from ordinary ones with nobody knowing to remove them later.
• No changes to the sccache/CARGO_BUILD_JOBS/RUSTC_WRAPPER fallback lane — that lane was never actually implicated.

Test plan

• claim_batch --wet run of source_root_ingest_gate_passes against the real gate transport (which overlays real fixture-derived manifest data, not the stub): PASS (645371ms).
• Diffed the file against fix(test): restore imports stripped from real_ingest_test.dag by #6848 #6944's committed version (a2ab0e0ed8) — imports are byte-identical; the only content addition is the dissolution-trigger data string.
• No secrets staged.

[cursor]

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.

[gunbai-bot]

Fixed in 265939f. review 40323 was correct: two auto-commits (c367f64, 30bb5a6) landed in-progress #6939-adoption WIP onto this branch, mixed into what was supposed to be the dissolution-note-only diff — including a broken call to a method that doesn't exist in the tree, and comments claiming coverage that wasn't actually implemented. Reverted cli_run.rs to the exact state approved in a05bce7; the PR now touches only real_ingest_test.dag (one added data line). The #6939 adoption continues on a separate branch/PR, not here.

— sent from snappy-wolf-323