consolidate vs consolidate-memory: the names are confusable enough that a handoff briefed the wrong one

[jsirish]

consolidate (session logs) and consolidate-memory (auto-memory files) have adjacent names and
adjacent-sounding scopes. A handoff written by one session briefed the next to "run consolidate"
for what was actually a memory-file job; the wrong skill was invoked and caught itself at its own
Phase 1, which is the good outcome but still a wasted cycle.

This is the second instance of the pair being confused — #36 (closed) was the same confusion from the
other direction, where consolidate's own body referenced a consolidate-memory that wasn't
installed.

The confusion is structural, not a one-off: the two ship from the same repo, only one of them is in
the Claude Code plugin (v0.12.0 ships consolidate/handoff/onboard; consolidate-memory is
npx-installed only), and nothing in either name says "logs" or "memory".

Ask — cheapest first:

1. Make the first line of each description say which store it touches, in those words.
2. Have consolidate detect a memory-shaped request at Phase 0 and redirect, rather than at Phase 1
   after reading.
3. Consider whether handoff should avoid naming a consolidation skill in its "next session" output
   at all, and say "run a consolidation pass" instead — the handoff is where the wrong name got
   written down.