…t flake

**Andon**: main has been red since 96d7349 (PR #869 merge). The failing
test is `quantize::tests::tests_25::test_f203_simd_faster_than_scalar_q4_0`:
single-shot timing on a 256×256 Q4_0 matvec got Scalar=122ms, SIMD=136ms
(speedup 0.90×) — not a regression, pure OS/CPU jitter.

**Root cause**: the test measured exactly one 100-iteration run of each
path. On shared CI runners, a single run is dominated by cache state,
frequency scaling, and neighbor-process preemption. SIMD timing was
sometimes slower than scalar purely from environmental noise.

**Fix**: warmup round + best-of-5 rounds, take the minimum of each. The
minimum is a lower-jitter estimator of the underlying hardware cost. If
SIMD's best measurement is still slower than scalar's best, that's a
real regression worth failing CI — the Popperian falsification property
of F-203 is preserved, not weakened.

**Verification** (4090 Yoga runner, debug build):
  F203: Q4_0 Performance Falsification (best-of-5)
    Scalar (min): 47.93ms
    SIMD   (min): 46.58ms
    Speedup: 1.03x

Threshold `speedup > 1.0` unchanged. Test is now deterministic within
measurement precision.

Also picks up a pre-existing trailing-blank-line fmt drift in
`crates/aprender-serve/src/contract_gate.rs` that `cargo fmt -p
aprender-serve` corrected as a collateral effect.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

…t flake (#875)

**Andon**: main has been red since 96d7349 (PR #869 merge). The failing
test is `quantize::tests::tests_25::test_f203_simd_faster_than_scalar_q4_0`:
single-shot timing on a 256×256 Q4_0 matvec got Scalar=122ms, SIMD=136ms
(speedup 0.90×) — not a regression, pure OS/CPU jitter.

**Root cause**: the test measured exactly one 100-iteration run of each
path. On shared CI runners, a single run is dominated by cache state,
frequency scaling, and neighbor-process preemption. SIMD timing was
sometimes slower than scalar purely from environmental noise.

**Fix**: warmup round + best-of-5 rounds, take the minimum of each. The
minimum is a lower-jitter estimator of the underlying hardware cost. If
SIMD's best measurement is still slower than scalar's best, that's a
real regression worth failing CI — the Popperian falsification property
of F-203 is preserved, not weakened.

**Verification** (4090 Yoga runner, debug build):
  F203: Q4_0 Performance Falsification (best-of-5)
    Scalar (min): 47.93ms
    SIMD   (min): 46.58ms
    Speedup: 1.03x

Threshold `speedup > 1.0` unchanged. Test is now deterministic within
measurement precision.

Also picks up a pre-existing trailing-blank-line fmt drift in
`crates/aprender-serve/src/contract_gate.rs` that `cargo fmt -p
aprender-serve` corrected as a collateral effect.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

…on (#869)

The base FALSIFY-MCP-002 only asserts that each tool's `inputSchema.type ==
"object"`, which is necessary but not sufficient: the MCP spec requires the
schema to validate against JSONSchema Draft 7. A tool definition could ship
with a malformed `properties` map or a nonexistent `type` and still pass.

This adds a separate `tests/falsify_schema.rs` that compiles every shipped
tool's `inputSchema` with `jsonschema::validator_for`. Compilation performs
meta-schema validation as a side effect, so any malformed schema fails the
test before it can ship to a real client.

Includes a guard test that feeds a known-bad schema (`properties` set to a
string) to the same validator path, asserting that compilation rejects it.
This catches a future jsonschema upgrade that silently accepts garbage.

Dev-dep only: jsonschema 0.28 is already in Cargo.lock via aprender-train,
so no new transitive crates are pulled in.

Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>

…t flake (#875)

**Andon**: main has been red since 96d7349 (PR #869 merge). The failing
test is `quantize::tests::tests_25::test_f203_simd_faster_than_scalar_q4_0`:
single-shot timing on a 256×256 Q4_0 matvec got Scalar=122ms, SIMD=136ms
(speedup 0.90×) — not a regression, pure OS/CPU jitter.

**Root cause**: the test measured exactly one 100-iteration run of each
path. On shared CI runners, a single run is dominated by cache state,
frequency scaling, and neighbor-process preemption. SIMD timing was
sometimes slower than scalar purely from environmental noise.

**Fix**: warmup round + best-of-5 rounds, take the minimum of each. The
minimum is a lower-jitter estimator of the underlying hardware cost. If
SIMD's best measurement is still slower than scalar's best, that's a
real regression worth failing CI — the Popperian falsification property
of F-203 is preserved, not weakened.

**Verification** (4090 Yoga runner, debug build):
  F203: Q4_0 Performance Falsification (best-of-5)
    Scalar (min): 47.93ms
    SIMD   (min): 46.58ms
    Speedup: 1.03x

Threshold `speedup > 1.0` unchanged. Test is now deterministic within
measurement precision.

Also picks up a pre-existing trailing-blank-line fmt drift in
`crates/aprender-serve/src/contract_gate.rs` that `cargo fmt -p
aprender-serve` corrected as a collateral effect.

🤖 Generated with [Claude Code](https://claude.com/claude-code)