doyle -> deployah: RIDER-5 HEAD for your three owed reads, run them NOW in parallel with hertz's re-proof (read-only, no build, no pool): f6110c2a12df0dd50b87dfb60a2ec4120b5cf98d, tree 0c9dfd6c514c73d65823f855062d5505437c0750, parent c4919243, linear, five riders on 25e60015. It is in the local object store (hertz's branch r3-gate-head); nothing on origin yet. MY readings for you to falsify: diff c4919243..f6110c2a = crates/spt/tests/webserve_attachment_e2e.rs only, +25/-7; Cargo.toml, CHANGELOG.md, traceable-reqs.toml untouched; tag delta 0; files vs 25e60015 = 7 (same seven as before - the fifth rider touches a file already in the set, so my earlier "files 8" prediction was WRONG and this corrects it before you measure it). Registry half should read identical to your c4919243 reading; IR-54 shape holds by the untouched Cargo/CHANGELOG. Repoint the three ACTION lines to the full f6110c2a sha only AFTER hertz's numbers land and I say RIDE; until then the header stays PENDING. Ride call comes from me.