Point two pilot-table intake files at where those factories live now

[andrewboudreau]

main is red on tools/check_dead_references.py, and has been since #2127 landed:

FAIL: 2 prose reference(s) name a path that does not exist:
  notes/tu-filename-reconstruction-pilot.md
      names `src/FlyGuy_Spawn.c`, which is not in the tree
  notes/tu-filename-reconstruction-pilot.md
      names `src/WaterfallMist_Spawn.cpp`, which is not in the tree

Reproduced on a clean checkout of 3d4377d61, so this is not a branch artifact.

Neither PR was wrong on its own. #2127's pilot table was written against the tree as it
stood, and the TU promotions that deleted those two files were in flight beside it — two
green PRs, one red main. Every open PR now inherits the failure through its
merge-with-main check; it is what turned #2148 red, and #2130, #2151 and #2156 are
behind it too.

The fix

The table's second column is headed "Current intake file", so the honest fix is to
name the file each factory lives in today rather than to bank the dead paths in
config/dead-reference-baseline.json:

profile	was	is now
WATERFALL	src/WaterfallMist_Spawn.cpp	src/actors/d_a_obj_waterfall.cpp
PROPELLER_HEYHO	src/FlyGuy_Spawn.c	src/actors/daPropeller_Heyho_c.cpp

Both were checked by grep, not assumed — WaterfallMist_Spawn is defined in the waterfall
TU, and FlyGuy_Spawn is ROM ordinal 26 in daPropeller_Heyho_c.cpp. Every other row in
the table still names a live file.

Prose only; no build input is touched. check_dead_references passes on the branch.

🤖 Generated with Claude Code

https://claude.ai/code/session_01FsXWgEM3Gtbqs6zEdwLMYV

[tangos-validator]

✅ PR validation — Passed

noverify: no source/build-data changes in this PR

_{Each changed src/*.c|*.cpp is compiled and its relocated bytes compared to the binary data on a private build box. Passing requires every changed file to reproduce the ROM byte-for-byte with correct relocation targets — this catches WRONG-DEST relocations and non-reproducing near-misses that ledger-scoped linkcheck skips.}

[andrewboudreau]

Closing as an exact duplicate of #2159: same base, same file, and byte-for-byte identical two-line path repair. #2159 is older and already has all public checks green.