This is the single source of truth for all ecosystem metrics. Every other doc references this file. No doc should hardcode counts. Re-verify before citing — numbers change with every deploy.
| Metric | Command | Notes |
|---|---|---|
| MCP tools | curl mcp.holoscript.net/health → tools |
Production count |
| Compilers | find packages/core/src -name "*Compiler.ts" -not -name "CompilerBase*" -not -name "*.test.*" | wc -l |
Sovereign + bridge |
| Export targets | ExportTarget type in packages/core/src/compiler/CircuitBreaker.ts |
Enum members |
| Sovereign (native) targets | SOVEREIGN_TARGETS in packages/core/src/compiler/sovereign-targets.ts |
Targets HoloScript runs/renders itself (no third-party engine); SSOT for the D.006 split |
| Bridge targets | BRIDGE_TARGETS in packages/core/src/compiler/sovereign-targets.ts |
Targets that emit to a third-party engine/format/runtime |
| Native engines/renderers | SOVEREIGN_ENGINES in packages/core/src/compiler/sovereign-targets.ts |
Package-level sovereign engines (WebGPURenderer, snn-webgpu, holo-vm, …); see docs/native-engine-registry.md |
| Trait files | find packages/core/src/traits -name "*.ts" -not -name "*.test.*" | wc -l |
Individual definitions |
| Trait inventory (Paper 11) | pnpm --filter @holoscript/core exec vitest run src/traits/constants/__tests__/paper-trait-inventory.test.ts → .bench-logs/paper-trait-inventory.json |
Do not hardcode counts in TeX; cite JSON + benchmark-results-2026-04-18-paper-11-trait-harnesses.md |
| Trait conflict census (Paper 11) | pnpm --filter @holoscript/core exec vitest run src/traits/constants/__tests__/paper-trait-conflict-census.test.ts → .bench-logs/paper-trait-conflict-census.json |
Pairwise mode counts; see same benchmark markdown |
| Trait categories | ls -d packages/core/src/traits/*/ |
Category directories |
| Domain plugins | ls -d packages/plugins/*/ | grep -v template | wc -l |
Excludes template |
| Packages | ls -d packages/*/ | wc -l |
Workspace packages |
| Tests | pnpm test → summary line |
Expensive — cache CI output |
| Public floor claims (marketing) | README.md "Facts at a glance" · services/holoscript-net-v2/public/llms.txt · services/holoscript-net-v2/holo-pages/page.holo @metadata |
Floor-rounded "N+" claims for humans/AI crawlers — the ONE sanctioned hardcoded form. Growth never breaks a floor; before lowering any count, re-verify every floor against the matching command above (compilers, traits, targets) |
| TET10 GPU-CG vs CPU-CG | pnpm exec tsx scripts/tet10-gpu-cg-benchmark.ts --out .bench-logs/tet10-gpu-cg-benchmark.json |
Hardware-specific; cite the artifact path and adapter info before using the speedup |
| GPU solver paths | pnpm exec tsx scripts/gpu-solver-benchmark.ts --out .bench-logs/gpu-solver-benchmark.json |
Covers TET4 CSR GPU-CG plus thermal/acoustic RegularGrid3D stencil kernels; cite artifact path before using speedups |
| Knowledge entries | curl mcp-orchestrator-production-45f9.up.railway.app/health → knowledge_entries |
Orchestrator |
| R3F renderers | ls packages/r3f-renderer/src/components/ | wc -l |
Production components |
| Absorb tools | curl absorb.holoscript.net/health → tools |
May be down — check first |
| npm packages | npm search @holoscript 2>/dev/null | wc -l |
Published @holoscript/* + create-holoscript |
| Scaffold CLI source package | npm view create-holoscript version deprecated |
Historical npm name sourced from packages/create-holoscript; re-check deprecation before recommending |
| Scaffold CLI public quickstart | npm view create-holoscript-app version deprecated |
Current public quickstart name; compare with create-holoscript before claiming release parity |
| PyPI package | pip3 show holoscript |
Core + 6 domain bridge extras (medical, alphafold, astronomy, robotics, scientific, all) |
| Store packages | curl store.holoscript.net/-/v1/search?text=holoscript → total |
Verdaccio registry |
- Active docs (
docs/,README.md,FULL_README.md) must never hardcode ecosystem counts. Reference this file or the verification command. - Archive and historical docs (
docs/archive/,docs/_archive/, dated reports, reviews, examples-health snapshots, planning/vision docs, andCHANGELOG.md) stay frozen or date-qualified as point-in-time records. - Marketing docs (
docs/marketing/) may snapshot numbers with a verification date:"158 MCP tools (verified 2026-04-13)". Must re-verify before posting. Files explicitly marked archived/stale are not reusable campaign copy. - Strategy docs (
docs/strategy/) may use counts if dated. Live battlecards should preferdocs/NUMBERS.mdor a verification command for HoloScript's current counts. The ROADMAP.md is authoritative for version-specific milestones. - Demo files (
docs/demos/) should reference live endpoints where possible.
Every hardcoded number becomes a lie within weeks. The MCP server had 34 tools, then 82, then 88, then 122, then 158 — and every doc that pinned a number is now wrong. The compilers went from 15 to 24 to 28 to 44. The traits went from 1,800 to 3,300+. Tests went from 2,460 to 44,000+.
The fix is not "update the number." The fix is "stop hardcoding the number."