Back to registry
⌁ Mapped Project • Mapped, proof pending Updated Not dated

AI Swarm Simulator

A source-independent executable Buildprint remapped from MiroFish as a product/outcome-flow AI swarm simulator. It now requires a real implementation-project foundation before Phase 01: AGENTS.md, architecture.md, engineering-standards.md, test-strategy.md, ui-identity.md, and a base project structure that coding agents must follow.

What this builds

A scoped product implementation packet.

Build a production-grade AI swarm simulator without depending on the original source repo. Status remains PROOF_REQUIRED until all proof gates pass.

  • Executable Buildprint packet with BUILDPRINT.md canonical start
  • Foundation scaffold gate requiring AGENTS.md, architecture.md, engineering-standards.md, test-strategy.md, and ui-identity.md
  • Five outcome-flow phases covering seed/ontology, graph memory, simulation environment, simulation run observability, and report interaction
  • Role contracts for product architecture, backend services, frontend workbench, and QA/security
  • Generated agent prompt and seed evidence ledger with claim status kept PROOF_REQUIRED
Core capabilities

The useful parts the finished build should expose.

01 Executable Buildprint packet with BUILDPRINT.md canonical start
02 Foundation scaffold gate requiring AGENTS.md, architecture.md, engineering-standards.md, test-strategy.md, and ui-identity.md
03 Five outcome-flow phases covering seed/ontology, graph memory, simulation environment, simulation run observability, and report interaction
04 Role contracts for product architecture, backend services, frontend workbench, and QA/security
05 Generated agent prompt and seed evidence ledger with claim status kept PROOF_REQUIRED
What you need

Local first, live proof explicit.

  • A clean implementation workspace
  • LLM provider credentials for live ontology/report/profile proof, or acceptance that only live proof remains blocked
  • Zep-compatible graph-memory credentials for live graph proof, or a configured adapter/test-double path
  • OASIS/CAMEL-compatible runtime or approved simulation worker, plus deployment authorization for production proof
  • Persistence/backend and browser test environment capable of repeatable e2e proof
System shape

What kind of artifact this becomes.

Product surface

Build a production-grade AI swarm simulator without depending on the original source repo. Status remains PROOF_REQUIRED until all proof gates pass.

Runtime layer

Full-stack web runtime, Foundation scaffold gate, Provider adapters, Worker/runtime lifecycle, Browser/e2e proof required

Build materials

Foundation scaffold / Seed intake / Ontology / Graph memory

Proof boundary

Executable packet / Foundation scaffold required

Build scope

Included, required from you, and outside the claim.

Included
  • Executable Buildprint packet with BUILDPRINT.md canonical start
  • Foundation scaffold gate requiring AGENTS.md, architecture.md, engineering-standards.md, test-strategy.md, and ui-identity.md
  • Five outcome-flow phases covering seed/ontology, graph memory, simulation environment, simulation run observability, and report interaction
  • Role contracts for product architecture, backend services, frontend workbench, and QA/security
Bring yourself
  • A clean implementation workspace
  • LLM provider credentials for live ontology/report/profile proof, or acceptance that only live proof remains blocked
  • Zep-compatible graph-memory credentials for live graph proof, or a configured adapter/test-double path
  • OASIS/CAMEL-compatible runtime or approved simulation worker, plus deployment authorization for production proof
Out of scope
  • Overclaiming MiroFish clone parity
  • Counting deterministic adapters as live provider proof
  • Treating missing credentials as permission to omit provider adapters/config/tests
  • Static graph/report data presented as real behavior
Agent handoff

Start from the packet, not the UI.

agb start https://agent-buildprint.com/buildprints/ai-swarm-simulator/package.json
Key files

The first files an agent should read.

All package files
01-questions.md Buildprint package file
02-project-setup.md Buildprint package file
03-phases/01-seed-ontology-workbench.md Buildprint package file
03-phases/02-graph-memory-build.md Buildprint package file
03-phases/03-simulation-environment.md Buildprint package file
03-phases/04-simulation-run-observability.md Buildprint package file
03-phases/05-report-and-interaction.md Buildprint package file
03-phases/phase-flow.md Buildprint package file
03-phases/phase-index.yaml machine-readable mirror
04-evaluation.md Buildprint package file
05-evidence/evidence-ledger.jsonl Buildprint package file
05-evidence/evidence-ledger.schema.json machine-readable mirror
06-contracts/data-persistence.md Buildprint package file
06-contracts/integration-runtime.md Buildprint package file
06-contracts/product-architect.md Buildprint package file
06-contracts/security-boundary.md Buildprint package file
06-contracts/test-and-verification.md Buildprint package file
06-contracts/ux-ui-craft.md Buildprint package file
blueprint.yaml machine-readable executable packet router
BUILDPRINT.md compatibility bootstrap or package contract
checks/acceptance.md acceptance checklist
CONTRACTS.md legacy interface/data contracts, when present
generated/agent-prompt.md Buildprint package file
PLAN.md legacy execution index, when present
publication.json machine-readable mirror
README.md human overview, non-authoritative
SPEC.md legacy behavior requirements, when present
TEST_MATRIX.md legacy risk-to-test alignment, when present
VALIDATION_TEMPLATE.md legacy completion report template, when present