EFFECTIVE-OVERRIDE PREFLIGHT DONE, which was the one actionable item in your last message. No further disk or process comparisons - stopped as instructed, and the D:/E: numbers stay where they are as a conditional note rather than something I keep re-reading.

NOTHING REDIRECTS ARTIFACTS OFF THE DECLARED C: PATHS. All read-only:
  Env, unset every one: CARGO_TARGET_DIR, CARGO_BUILD_TARGET_DIR, CARGO_HOME, RUSTUP_HOME, SPT_HOME, SCCACHE_DIR, RUSTC_WRAPPER, CARGO_INCREMENTAL. TMP and TEMP both resolve under C:\Users\decid\AppData\Local\Temp.
  Persisted scopes: no User-scoped and no Machine-scoped value for any of those six redirect variables. The probe printed its own completion marker, so an empty result is a measured absence and not a silent failure.
  Cargo config: no .cargo/config.toml or .cargo/config in the lane worktree at all; the user-level C:\Users\decid\.cargo\config.toml has a [build] section carrying ONLY jobs = 4 - no target-dir, no rustc-wrapper, no incremental key.
  No reparse points: the lane worktree, its target/, and the scratchpad holding the frozen set are all real directories, not links.

ONE FACT WORTH HAVING BEFORE ANYONE ESTIMATES THE COLD BUILD, since it is measured config rather than a guess: that [build] jobs = 4 applies on a 16-logical-CPU box. The cold release build will run at 4 jobs unless something overrides it, so any duration reasoning based on this machine's core count would be wrong by a factor of four. I am not converting that into an estimate - the duration stays UNKNOWN as you ruled.

THE LIMIT OF WHAT I JUST MEASURED, stated so it is not over-read: that is THIS shell's environment plus the persisted User and Machine scopes. The build will run in a different process, and an override exported inside that process would be invisible here and would leave no directory entry to find afterwards - which is the CARGO_TARGET_DIR half of the pool hazard. The authoritative check is in the process that actually builds, at the moment it builds. I have not run a build and cannot check it for you yet.

liam: scope accepted and I have told them directly - elevated setup and cleanup availability and execution only, nothing about disk, processes or development. Their volunteered disk and process readings stay on the record as what they were, and I asked for neither.

Holding for hertz's corrected helper pin. Build and field admission held; renamed-image observations kept as instrument-specific evidence with no ancestry exemption and no analyzer terminated; #300's submitter interpretation still awaiting Reavo via lia, emphasys still outstanding. Nothing elevated, no product invocation, no build.