BISECT: drop only the profile debug change to isolate the E2E failure

Not for merge. T1's E2E lane has failed on all seven runs while T2/T3/T4 from
the same base are green. This reverts ONLY [profile.dev] debug = 0 (and the
release restatement), leaving every other T1 change in place, so one CI run
says whether the profile block is responsible.

Green here => the profile change causes it.
Red here   => it is elsewhere in T1 and this hypothesis is dead.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Cargo.toml

@@ -520,26 +520,7 @@ path = "tests/integration/run_artifact_timings.rs"
 name = "reborn_integration_db_write_canonical"
 path = "tests/integration/db_write_canonical.rs"
 
-# Debug-info policy for dev and test (test inherits dev, so this one line
-# governs `cargo test` binaries and their dependencies too). CI has built
-# every heavy lane with CARGO_PROFILE_DEV_DEBUG=0/CARGO_PROFILE_TEST_DEBUG=0
-# for a long time; codifying it here makes local builds reproduce CI and lets
-# the per-workflow env copies be deleted as no-ops. Assert/panic messages
-# keep their file:line via panic::Location at any debug level; debug=0 only
-# drops mid-frame file:line from RUST_BACKTRACE output. Measured 2026-08-21
-# on ironclaw_common tests (rustc 1.98.0): debug=0 175 MiB / line-tables-only
-# 257 MiB (+47%) / full 272 MiB — the smaller artifacts win on 2-core CI
-# runners and the 10 GB cache budget. Debugging locally with variable
-# inspection? Override per-run: `CARGO_PROFILE_DEV_DEBUG=2 cargo <cmd>`.
-[profile.dev]
-debug = 0
-
 [profile.release]
-# Already cargo's built-in default (debug info is off by default in
-# release); made explicit because ironclaw-stress.yml restated it three
-# times as CARGO_PROFILE_RELEASE_DEBUG=0 workflow/job env — this line makes
-# those deletable no-ops too (Task 9).
-debug = 0
 strip = true          # Remove debug symbols from release binaries
 
 # The profile that 'cargo dist' will build with