Add API benchmark suite

[s0584273828-ctrl]

/claim #30

Summary

• Add a dependency-light API benchmark suite under benchmarks/ covering /health and every mounted /api/* route.
• Add realistic synthetic payloads, generated benchmark admin auth, threshold gates, JSON output, and Markdown output.
• Add root commands: npm run benchmark and npm run benchmark:smoke.
• Add a GitHub Actions smoke gate at .github/workflows/api-benchmark-smoke.yml.
• Add short demo proof at demos/api-benchmark-smoke-demo.mp4.

Smoke benchmark summary

Committed full report: benchmarks/results/latest.md and benchmarks/results/latest.json.

Endpoint coverage	Requests each	Error rate	Gate
21 endpoints (/health + all /api/*)	3	0%	pass

Slowest smoke p99 in the committed run: health at 15.13 ms; upload p99 was 2.09 ms.

Benchmark Environment

Hardware

• CPU model & core count: Apple M5, 10 cores reported by Node.js
• RAM: 32 GB total, about 5.2 GB free before implementation
• Storage type: local Apple SSD
• Network interface: loopback for the benchmark target
• Machine type: local workstation
• OS & version: macOS 26.4.1

Runtime

• Node.js version: v24.14.1
• Resource limits: no explicit benchmark resource limits
• Other significant processes: normal Codex desktop and shell background activity

AI-assisted execution disclosure

• Agent/tool: OpenAI Codex desktop
• Model: GPT-5-based Codex model
• Provider: OpenAI
• Orchestration framework: none beyond Codex shell/GitHub tooling
• Execution mode: human-initiated, agent-executed
• Shell/tool access: yes
• Internet access: yes, through approved GitHub/network commands
• Benchmark commands: run directly by the agent against local loopback
• Known constraints: sandboxed commands cannot bind sockets without approved escalation, so benchmark validation used approved local loopback execution.

Validation

• npm run benchmark:smoke -> all 21 endpoint gates passed
• GitHub Actions API benchmark smoke / smoke -> passing
• node --check benchmarks/run-api-benchmarks.mjs
• node --check benchmarks/scenarios.mjs
• ffprobe demos/api-benchmark-smoke-demo.mp4 -> 1280x720, 8s, 200 frames
• git diff --check

Prepared with AI assistance and manually reviewed before submission.

[s0584273828-ctrl]

Follow-up: added a GitHub Actions smoke workflow for the API benchmark suite after refreshing the token with the required workflow scope.\n\nValidation:\n- npm run benchmark:smoke\n- git diff --check