Skip to content

PLATFORM_SCALE.md

Latest commit

 

History

History
80 lines (57 loc) · 3.77 KB

PLATFORM_SCALE.md

File metadata and controls

80 lines (57 loc) · 3.77 KB
Raw

AgentStack platform scale (public facts)

As of: 2026-08-20 · Core line: 0.4.18
Use for: LinkedIn, marketplace listings, investor one-liners, plugin copy — do not hand-edit counts; refresh via the platform stats codegen chain (docs.freshness.living.gen1).

Russian edition: PLATFORM_SCALE_RU.md

Maintainers (monorepo SoT): docs/publication/platform-stats.snapshot.json on the default branch.

Headline numbers

Metric Value Notes
MCP catalog actions 656 GET /mcp/actions / MCP_CAPABILITY_MATRIX.md Total actions
Action domains 57 Domain sections in capability matrix (codegen)
MCP tools registered 695 MCP_TOOLS_REGISTERED = len(MCP_TOOLS_REGISTRY) at codegen — not catalog actions
IDE entry tool 1 agentstack.execute — batch steps, discovery via /mcp/actions
Plugin surfaces 4 Cursor, Claude Code, GPT, VS Code

Marketing shorthand (accurate): agent actions · domains · one protocol for humans and AI.

MCP count terminology

Metric Example Use
Catalog actions (public) 656 Marketing, GET /mcp/actions, MCP_CATALOG_ACTIONS
Registry tools (runtime) ~695 GET /mcp/health → tools_count; startup log mcp_registry_tools
MCP tools registered (constant) 695 MCP_TOOLS_REGISTERED — synced at codegen; not catalog actions

Never rewrite SEO copy from tools_count alone (MET-01).

Top domains (by action count, Jul 2026)

Domain Actions Typical use
social 83 Messenger, feeds, channels, presence
integrations 48 Integration Hub, webhooks, recipes
agentnet 43 AgentCoin ledger, bridge, checkpoints, rails
commerce_rest 39 Marketplace / commerce REST catalog hints
agents 25 Agents Fleet CRUD, runs, policy
logic 19 Logic Engine rules and execution
commerce 18 Commerce kit MCP actions
hosting 18 Sites, publish pipeline, buckets
bots 17 Bots Fleet
crm 17 CRM tissue
projects 14 Project lifecycle, users, stats
scheduler 12 Tasks, pools, cron-style jobs

Full matrix: MCP_CAPABILITY_MATRIX.md (public mirror; redacted).

How to verify (production or staging)

GET https://agentstack.tech/mcp/health
GET https://agentstack.tech/mcp/actions

Expect catalog totals consistent with this sheet after deploy.

Refresh checklist (maintainers)

  1. Regenerate the MCP capability matrix via the platform stats codegen chain (internal freshness runbook on the default branch).
  2. node scripts/codegen-platform-stats.mjs
  3. Align shared/constants.py MCP_CATALOG_ACTIONS / MCP_ACTION_DOMAINS
  4. Update this file + PLATFORM_SCALE_RU.md (use <!-- stats:* --> for counts and marketing shorthand lines)
  5. node scripts/patch-public-doc-stats.mjs
  6. npm run sync:mcp-marketing-prose (fundraising, grants, decks, PUBLISHER_COPY, core_app.py OpenAPI blurbs)
  7. npm run audit:stats-plane-parity and npm run audit:hardcoded-mcp-stats:widen

Genetic tag: docs.publication.platform_scale.gen1 · Living Plane: docs.freshness.living.gen1