test(cli): kill the surviving mutants of the framework context

[blafourcade]

Problem

The framework mutation scope holds at its floor of 69 with 1235 surviving and 383 uncovered mutants of 5646, the largest hole in the tree: contexts/framework is 88 files, 38 % of src/, and a third of its mutants live. Every other context now sits between 85 and 97 after #795; a floor of 69 means a behaviour change in install, sync or restore can pass the suite.

Scope

• Per use case, read the survivors in reports/mutation/framework/ and write the test each one names, red first against the mutation, then restore.
• Prefer whole-structure assertions (toStrictEqual on the result, exact messages) over presence checks; a Set or Map never goes inside expect.objectContaining.
• Raise the floor to the measured score minus two.
• Survivors argued equivalent are listed in the PR with their line and reason.

Acceptance criteria

• node scripts/run-mutation.mjs framework --force ≥ 85
• Every new it shown red first, one line of evidence per test in the PR
• No source change, or each one named and justified in the PR
• Every gate green

Prior art in this repo

• feat(cli): one shared framework source per machine, user scope and session telemetry #795: the same pass over presentation (74.0 → 97.3), translate (68.0 → 87.0) and the five tool profiles (44–65 → 80–94), in cli/tests/presentation/commands/*-wiring.integration.test.ts and cli/tests/contexts/tools/domain/profiles/*/build.unit.test.ts.
• refactor(cli): make the refactor's guarantees provable — determinism, mutation, golden, dead code #750, "Out of scope": plugin-fetcher-adapter, self-update/git-adapter, auth-provider-adapter roughly 105 uncovered mutants, phase 5 of 2026_09_03_mutants-sans-couverture/, never written.

Out of scope

• Splitting contexts/framework into install, sync, restore. A design question wanting a brainstorm; the tests written here should survive it.