fix(coord): -Take and -Release recommended -Force on a holder nobody had looked at (BACKLOG #345 Half B)

[wshallwshall]

Completes BACKLOG #345. Half A landed in #141; this is the other half.

What

#106 taught -List to probe whether a claim's holder still exists on disk. That was the wrong half to fix alone: -List is where you browse, -Take and -Release are where a session is stopped and has to choose between waiting, picking other work, and taking the key.

Both blocking paths printed the same "held by another worktree" block whether the holder had been deleted, had died, or was committing that minute. -Release went further and printed:

If that session is gone, re-run with -Force.

— unconditionally, about a holder it had never probed. That's an instruction to guess, issued at the exact moment the guess is expensive: the cheapest way past the gate was also the one that frees a live session's key. claim.ps1's own header records the cost — three sessions fixed the same npm advisory on 2026-07-24, two PRs closed as duplicates, and the survivor hadn't tested what the others found.

The change

All three surfaces now share one liveness helper, because they had been disagreeing — -List probed, the other two didn't probe at all.

Holder state	-Take blocked	-Release refused
gone	names the exact take-over commands	recommends -Force
present	withholds the -Force recipe, says quiet is not dead	says ask that session first
unknown / probe failed	treat as live, coordinate	confirm before -Force

The asymmetry is the design

A vanished worktree is a fact and the one state safe to act on unasked. Present, undatable and unprobeable all read as "coordinate first", never "probably fine" — a session can be alive and simply not committing, so silence is not evidence of death. The catch returns failed, never gone: an unreadable path must not become a licence to release someone's live claim.

A probe hardwired to gone would pass every positive assertion, so each positive case is paired with the negative one that catches it.

-Force itself is deliberately untouched. This reports; it does not enforce. Refusing to override a live claim would strand every key whose holder is merely unreachable — the same orphan, one level up.

Verification

• 4 of 9 new tests fail against main's claim.ps1; the other 5 are regression guards that must pass on both sides (including two pinning fix(coord): claim -List measured age and called it staleness #106's -List behaviour through the refactor).
• One assertion was rewritten mid-build after it failed on the patched code — it searched for the token -Force, which the live-holder text legitimately contains inside "do NOT -Force it". The property wanted is "no runnable recipe", so it now asserts absence of the command line. Re-checked against unpatched afterwards to prove the rewrite hadn't made it vacuous.
• 41 coordination tests green; ruff clean; no non-ASCII introduced into a BOM-less script.

Sequencing note

Deferred once on purpose: claim.ps1 had three sessions in it at once (#106, #140, this). Built once both merged — one round-trip, no rework, no three-way merge.

🤖 Generated with Claude Code