enhance(#3619): ratchet the platform seam with local/no-private-binary-resolution (epic #3411 Phase 3) (#3636)

* chore(#3619): ratchet the platform seam with local/no-private-binary-resolution

Epic #3411 Phase 3, the ratchet. Scope revised with maintainer approval and
recorded on the issue: the epic's literal ask was a rule rejecting a bare-name
spawn outside the seam. Surveyed at ac1b6d679, ~30 such sites exist and none is
a defect — git, gh and npm ship native .exe that CreateProcess resolves unaided,
and the rest are POSIX-only tools. ADR-1703 rules 2 and 3 forbid grandfathering
and escape hatches, so a literal rule would be unsuppressable and would force
rewriting 30 correct calls.

The epic's actual thesis was four private RESOLVERS, not four bare spawns. So
the rule flags re-implementing resolution: reading PATHEXT in any casing from
any object, and a hardcoded list carrying two or more of .exe/.cmd/.bat/.com —
precisely the shapes fallow-runner's candidateNames and gsd-tools' PATHEXT
string had before Phases 1 and 2 deleted them.

Three boundaries were arrived at rather than assumed:

  two-or-more   a single .endsWith('.cmd') is a classification, not a candidate
                set; runtime-hooks-surface derives .cmd shim paths that way
  boundary-aware  a naive substring test flags .execute and .compacting, caught
                on src/host-integration.cts before it could become a false
                positive nobody could suppress
  suffix-anchored  the seam exemption matches src/shell-command-projection.cts
                exactly; a substring match would also exempt the dispatch test
                file. Case I9 pins it.

PATH scans are deliberately NOT flagged — membership checks (bin/install.js)
are indistinguishable from resolution scans, and an unsound rule in a
zero-escape-hatch architecture is worse than no rule.

To make the ratchet strict with no carve-out, resolveExecutableBinary gained
pathOverride: search THIS PATH, read everything else including PATHEXT from the
ambient environment. resolveFallowBinary now supplies its own search path
without hand-threading PATHEXT, which would itself have been a private read.
The three alternatives were all worse: exempting the file is grandfathering,
exempting the AST shape is a carve-out every future caller must replicate, and
dropping the pass-through would silently ignore a user's real PATHEXT — buying
a lint rule with a correctness regression.

eslint-rules/** is outside the rule's globs rather than exempted, because
portability-vocab.cjs owns the extension set. scripts/**/*.cjs got its own block
so that exclusion does not leave a hole in the ratchet.

Started green with nothing suppressed. Proven able to fail: a fixture with both
signals reports two errors.

Refs #3411

* fix(#3619): close the PATHEXT destructuring evasion and correct two overclaims

Adversarial review found a trivial evasion of the rule's primary signal: the
visitor only handled MemberExpression, so

  const { PATHEXT } = process.env
  const { PATHEXT: exts } = process.env
  const { Pathext } = opts.env

were all unflagged. That is a common idiom, not an exotic bypass. An ObjectPattern
visitor now catches it in every form — renamed, any casing, any receiver, string
keys — while leaving a computed key alone, since it is not statically decidable.
I10-I13 pin the invalid forms and V9/V10 pin PATH and the computed key.

Two overclaims corrected, both mine:

Standards review proved the docs were factually wrong. Both the ADR amendment and
the CONTEXT.md entry asserted that tests/shell-command-projection-dispatch.test.cjs
is still linted by this rule. It is not — the rule's surface is src, gsd-core/bin,
scripts and hooks, and tests/** is deliberately outside it because test setup
legitimately assigns process.env.PATHEXT (fallow-runner's P3 does exactly that).
The suffix-vs-substring distinction is therefore proven by RuleTester case I9
feeding a synthetic filename, NOT by real coverage of that file. Both documents now
say so.

The rule's own docstring claimed the seam exemption matches the seam path
'exactly'. It is a suffix match, so a nested foo/src/shell-command-projection.cts
would also be exempt. Suffix matching is kept — it is how sibling rules resolve
paths and the nested case does not exist — but the docstring now states the
boundary rather than overstating the precision.

The evasion fix was verified by executing eslint against both destructuring forms
in scripts/, not by inspection. Probe: 31/31.

Refs #3411

* chore(#3619): backfill changeset pr number 3636

---------

Co-authored-by: sim <sim@local>

docs/adr/1703-portability-enforcement-architecture.md

@@ -76,6 +76,39 @@ Replace all three with a single coherent mechanism: **AST-based ESLint rules in
 | `require-userprofile-with-home` | `DEFECT.WINDOWS-TEST-PORTABILITY` (G6) | tests |
 | `normalize-path-in-content` | `DEFECT.WINDOWS-PATH-LEAK-IN-MARKDOWN-CONTENT` (`RULESET.CONTENT-PATH-NORMALIZATION`) | `src/**/*.cts` |
 | `require-fs-op-fallback` | `DEFECT.WINDOWS-FS-OPS` | `src/**/*.cts`, build/install |
+| `no-private-binary-resolution` | `DEFECT.WINDOWS-PRIVATE-BINARY-RESOLUTION` | `src/**/*.cts`, `gsd-core/bin/**`, `scripts/**`, `hooks/**` |
+
+**Amendment (2026-08-18, epic #3411 Phase 3 / #3619).** `no-private-binary-resolution` is the
+first catalog entry added after the original seven, and it extends this architecture to a
+**production** Windows-runtime class rather than a test-portability one. It flags the two
+unambiguous signals of re-implementing Windows binary resolution outside the platform seam:
+reading `PATHEXT` (in any casing, from any object), and a hardcoded list containing two or more of
+`.exe`/`.cmd`/`.bat`/`.com`. Both are exactly the shapes the four resolvers that epic #3411 deleted
+actually had.
+
+It deliberately does **not** flag a bare-name `spawn`, which is what #3411's own text asked for:
+~30 such call sites exist and none is a defect (`git`, `gh`, `npm` ship native `.exe` on Windows),
+so under rules 2 and 3 above a literal rule would be unsuppressable and would force rewriting
+correct code. It also does not flag a `PATH` scan, which is used for legitimate membership checks
+(`bin/install.js`) and cannot be soundly distinguished from a resolution scan. An unsound rule in a
+zero-escape-hatch architecture is worse than no rule.
+
+Two scoping consequences worth recording, because both were arrived at rather than assumed:
+
+- **`eslint-rules/**` is outside this rule's surface** — `lib/portability-vocab.cjs` owns the
+  extension set per rule 4 and would otherwise flag itself. This is expressed by *not linting that
+  tree*, never by an exemption, so rule 3 holds.
+- **The seam exemption is path-suffix-anchored**, matching `src/shell-command-projection.cts`
+  or any path ending in `/src/shell-command-projection.cts` — not a substring match. The rule's
+  own configured surface is `src/**/*.cts`, `gsd-core/bin/**/*.cjs`, `scripts/**/*.cjs`, and
+  `hooks/**/*.js`; `tests/**` is deliberately outside that surface, because test setup
+  legitimately assigns `process.env.PATHEXT` (`tests/fallow-runner.test.cjs`'s P3 case does this).
+  So the suffix-vs-substring distinction is proven only by case I9 of the rule's `RuleTester`
+  suite, which feeds the rule a synthetic filename directly — not by real-world linting of
+  `tests/shell-command-projection-dispatch.test.cjs`, which this rule never scans.
+
+The rule started **green** with nothing suppressed and nothing grandfathered — Phases 1 and 2 had
+already removed every private resolver, which is what made a strict ratchet possible at all.
 
 **Taxonomy coverage.** This catalog addresses every `DEFECT.WINDOWS-*` class plus
 `DEFECT.TEST-SHELL-PIPELINE-NONPORTABLE` in `CONTEXT.md`, to the extent each is *statically*