docs: add ModeVectorSet to hosted OpenAPI

[Code0987]

Summary

Fill the gap in the existing hosted docs (Swagger UI at https://code0987.github.io/supercache/ and http://<admin>/docs).

Cache OpenAPI listed ModeVectorSet in the mode table but had no VAdd / VRem / VSim / VCard / VDim / VEmb pages.

• Add those six RPCs + schemas on the Cache gRPC tab
• sc examples (sc -keyspace vectorset vadd …)
• Caps: dim 2–256, 512 members, metric cosine/l2/ip
• Matching section in docs/API.md
• KV examples use cacheonly instead of demo
• Spec version 0.13.0

Source of truth is api/openapi/ (embedded on the node). docs/api/ is the static copy. Pages deploys from api/openapi + pkg/admin/docsui on merge to main.

Release

• This PR should not cut a release

Test plan

• go test ./api/openapi ./pkg/admin
• After merge: GitHub Pages Docs workflow + /docs Vector tab on a node

Notes for reviewers

Docs only. No data-path change.

[github-actions]

SuperCache bench vs main

Both sides ran once on this same GitHub runner (main, then this PR). Not a merge gate.

Baseline: 830420d (2026-09-06T12:46:05Z) → this run: 830420d (2026-09-06T12:45:19Z)

scbench smoke

cell	ops/s (main → PR)	Δ ops/s	p50	p99
get/hit n=1 c=1 uniform	8591 → 8658	+0.8%	116µs → 116µs	194µs → 199µs
get/hit n=1 c=10 uniform	32926 → 33899	+3.0%	274µs → 265µs	840µs → 818µs
get/hit n=3 c=10 uniform	38639 → 38330	-0.8%	227µs → 229µs	904µs → 912µs
miss/miss-cacheonly n=1 c=1 uniform	8488 → 8503	+0.2%	116µs → 117µs	218µs → 213µs
miss/miss-cacheonly n=1 c=10 uniform	34779 → 34761	-0.0%	260µs → 259µs	840µs → 848µs
set/put n=1 c=1 uniform	8299 → 8120	-2.2%	118µs → 119µs	213µs → 292µs
set/put n=1 c=10 uniform	32190 → 32254	+0.2%	278µs → 278µs	878µs → 877µs

Δ ops/s: green-ish is higher throughput. Latency: lower is better.

go test -bench (ns/op)

benchmark	ns/op (main → PR)	Δ	allocs/op
BenchmarkEngineDelete	778 → 803	+3.2%	6 → 6
BenchmarkEngineGetHit	1211 → 1541	+27.3% worse	15 → 15
BenchmarkEngineGetHitParallel	1244 → 1371	+10.2%	15 → 15
BenchmarkEngineGetMissCacheOnly	913 → 906	-0.8%	14 → 14
BenchmarkEngineGetMissCacheOnlyParallel	686 → 655	-4.5%	14 → 14
BenchmarkEngineGetMissLoadThrough	1722 → 1622	-5.8%	24 → 24
BenchmarkEngineGetMissLoadThroughParallel	1973 → 1904	-3.5%	23 → 23
BenchmarkEnginePut	1667 → 1651	-1.0%	18 → 18
BenchmarkEnginePutParallel	1499 → 1440	-3.9%	18 → 18
BenchmarkStoreBitSet	112 → 113	+0.9%	1 → 1
BenchmarkStoreCGetHit	72 → 73	+0.2%	0 → 0
BenchmarkStoreCIncr	102 → 102	+0.4%	1 → 1
BenchmarkStoreCMSIncr	120 → 126	+4.7%	0 → 0
BenchmarkStoreCMSIncrN	120 → 121	+0.8%	0 → 0
BenchmarkStoreGeoAdd	9572 → 9577	+0.1%	2 → 2
BenchmarkStoreGeoPosHit	88 → 88	-0.0%	0 → 0
BenchmarkStoreGeoPosMiss	87 → 91	+4.3%	0 → 0
BenchmarkStoreGetHit	295 → 296	+0.5%	2 → 2
BenchmarkStoreGetHitParallel	348 → 379	+8.9%	2 → 2
BenchmarkStoreHGetHit	95 → 91	-4.3%	1 → 1
BenchmarkStoreHGetMiss	85 → 85	+0.6%	0 → 0
BenchmarkStoreHLLAdd	100 → 100	+0.2%	0 → 0
BenchmarkStoreHSet	10606 → 10600	-0.1%	3 → 3
BenchmarkStoreJsonSet	26294 → 26757	+1.8%	207 → 207
BenchmarkStorePut	304 → 311	+2.3%	2 → 2
BenchmarkStorePutParallel	357 → 358	+0.3%	2 → 2
BenchmarkStoreSetAdd	9289 → 9259	-0.3%	2 → 2
BenchmarkStoreSetContainsHit	87 → 88	+0.7%	0 → 0
BenchmarkStoreSetContainsMiss	84 → 84	+0.0%	0 → 0
BenchmarkStoreTopKAdd	1009 → 1105	+9.5%	20 → 20
BenchmarkStoreVAdd	109151 → 110927	+1.6%	512 → 513
BenchmarkStoreVSim	89605 → 104168	+16.3%	530 → 530
BenchmarkStoreZAdd	10002 → 9715	-2.9%	2 → 2
BenchmarkStoreZScoreHit	94 → 86	-8.0%	0 → 0
BenchmarkStoreZScoreMiss	86 → 87	+1.2%	0 → 0

_{Posted by CI bench job. Not a merge gate.}