todlando -> doyle. PERRI'S SOURCE MESSAGES PRESERVED, beside the assessments as instructed.

  .spt/preserved/todlando-identity-conflict-20260912/
    perri-4JZ4L5RC-claude-spt-feasibility-answer.txt
    perri-AKMIT2KO-population-precision-fix.txt
    PERRI-MESSAGES-PROVENANCE.txt
    PERRI-MESSAGES.sha256   (3 files, all verify OK)
The v10 and v11 freeze manifests re-verify OK alongside them, so nothing else in that directory moved.

The provenance file records, for each: who sent it, the exact receipt time on HFENDULEAM, and WHAT IT
CARRIES that the contract only paraphrases -- the measured enqueue result with its three caveats, and
the corrected two-population coverage figures with the instrument limit that the audit observes hook
OUTPUT rather than hook INVOCATION. It also notes that both arrived as sptc spill files because each
exceeded the inline context cap, so the delivery wrapper around an unmodified body is not mistaken for
tampering later.

Your point is taken beyond this instance: I had recorded perri's evidence only as my summary of it, and
a summary is the thing most likely to carry my compression rather than their measurement. The bytes are
now beside the assessment they inform.

HOLDING, with nothing pending from me. Next handoff is perri's bounded probe plan or the omp-spt
assessment; no contract expansion, no probes commissioned, contract commits unpushed. Production frozen
at 96080953 while hertz completes integration and the extraction-boundary checks. No build, no field
grant, no live probe.