fix: move benchmark schema to public dir and require result_kind

Applies the remaining code-review changes that the test suite already
expected but the source did not satisfy:

- Move results.schema.json into docs/public/benchmarks/ so VitePress
  serves it at /neuron-js/benchmarks/results.schema.json (matching its
  $id); previously it lived in a content dir and was never served and
  the contract test could not find it.
- Add result_kind to the per-row required array in the schema.
- Update path references in methodology.md and benchmarks/README.md, and
  fix the index.md link to the served public URL.

Fixes the failing CI quality job (3 benchmark-contract tests).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Author: sdieguez-wotc

benchmarks/README.md

@@ -4,6 +4,6 @@ This directory contains benchmark data artifacts for NJS-GROWTH-07.
 
 - `sample-results.placeholder.json` is deterministic placeholder data for chart and playground wiring.
 - It is not benchmark output and must not be used for public performance claims.
-- Real benchmark output must match `docs/benchmarks/results.schema.json` and set `result_kind` to `actual_benchmark`.
+- Real benchmark output must match `docs/public/benchmarks/results.schema.json` and set `result_kind` to `actual_benchmark`.
 
 See `docs/benchmarks/methodology.md` for the competitor, scenario, input-size, and metric contract.

docs/benchmarks/index.md

@@ -7,7 +7,7 @@ This section is a design and generation foundation. It does not publish benchmar
 ## Files
 
 - [Benchmark methodology and result contract](./methodology.md): competitor set, scenario matrix, input-size matrix, metric definitions, and placeholder-data policy.
-- [Result JSON schema](./results.schema.json): machine-readable contract for benchmark output and downstream chart data.
+- [Result JSON schema](/benchmarks/results.schema.json): machine-readable contract for benchmark output and downstream chart data.
 - [Visual proof system](./visual-proof-system.md): palette, typography, composition, diagram style, chart rules, social-card constraints, and README proof strip guidance.
 - [Visual proof prompt kit](./prompt-kit.md): reusable prompts for benchmark infographics, explainability trace diagrams, playground README GIF storyboards, AI-rule safety carousels, and README proof strips.
 - [Asset folder recommendation](./assets/): recommended `docs/benchmarks/assets/` structure and data-integrity rules.

docs/benchmarks/methodology.md

@@ -16,7 +16,7 @@ The harness/result contract covers exactly these engines for the first public pr
 | Hand-coded TypeScript | `hand-coded-typescript` | Baseline for direct conditional logic without engine overhead. |
 | rule-engine-js | `rule-engine-js` | Smaller modern competitor selected because it installs/builds in this repository. |
 
-`rulepilot` remains an alternate candidate only if `rule-engine-js` becomes infeasible later. Do not mix both in the same first chart set without updating `docs/benchmarks/results.schema.json`.
+`rulepilot` remains an alternate candidate only if `rule-engine-js` becomes infeasible later. Do not mix both in the same first chart set without updating `docs/public/benchmarks/results.schema.json`.
 
 ## Scenario matrix
 
@@ -37,7 +37,7 @@ The harness/result contract covers exactly these engines for the first public pr
 
 ## Stable result fields
 
-Every result row must contain these fields. Units and source definitions are duplicated in `docs/benchmarks/results.schema.json` for machine consumers.
+Every result row must contain these fields. Units and source definitions are duplicated in `docs/public/benchmarks/results.schema.json` for machine consumers.
 
 | Field | Unit | Source |
 | --- | --- | --- |
@@ -79,14 +79,14 @@ A future executable harness should:
 3. Execute `smoke`, `small`, and `medium` profiles by default; gate `large` behind an explicit flag.
 4. Measure cold start separately from warm throughput.
 5. Measure validation and explanation overhead as delta timings against the same scenario/input profile.
-6. Emit JSON matching `docs/benchmarks/results.schema.json`.
+6. Emit JSON matching `docs/public/benchmarks/results.schema.json`.
 7. Set `result_kind: "actual_benchmark"`, `is_placeholder: false`, and `claims_allowed: true` only for real measured output.
 
 ## Visual asset consumers
 
 The first visual asset bundle can safely bind to these paths:
 
-- Schema: `docs/benchmarks/results.schema.json`
+- Schema: `docs/public/benchmarks/results.schema.json`
 - Placeholder sample: `benchmarks/sample-results.placeholder.json`
 - Methodology: `docs/benchmarks/methodology.md`
 

docs/benchmarks/results.schema.json …cs/public/benchmarks/results.schema.jsondocs/benchmarks/results.schema.json renamed to docs/public/benchmarks/results.schema.json docs/benchmarks/results.schema.json renamed to docs/public/benchmarks/results.schema.json

@@ -70,7 +70,8 @@
           "explanation_overhead_ms",
           "node_version",
           "package_version",
-          "commit_sha"
+          "commit_sha",
+          "result_kind"
         ],
         "properties": {
           "engine": {