OntologyAI V6 — Decision Intelligence Platform

OntologyAI builds a live Ontology of a small business from its existing tools, and lets AI specialists query and act on it — with every consequential action gated by human approval. In V6 this becomes a living Enterprise Knowledge Model + stakeholder-specific Decision Workspaces, powered by observability-hardened mission pipelines.

Vision

Enterprise knowledge is scattered across meetings, docs, chat, spreadsheets, CRM, and ERP — creating duplicated effort, inconsistent decisions, and slow alignment. OntologyAI V6 converts that fragmented evidence into a living Enterprise Knowledge Model and stakeholder-specific Decision Workspaces that answer:

What is happening?
Why does it matter?
What evidence supports it?
What options exist?
What is recommended?
What are the trade-offs?
What happens if nothing changes?
Who owns the decision?
What changes downstream?

Outcome: One canonical model, many synchronized stakeholder views, and validated decision-ready artifacts.

Product Principle

Evidence first, knowledge second, decision third, artifact last.

No evidence → no artifact. No traceability → block. Low confidence → human review. Unresolved conflict → reject publish. Every layer depends on the layer before it.

Who it is for

Role	What they get
CEO / COO	Executive Decision Brief, Org / Ownership Map, Current-State Pack
CFO	Feasibility & Options Report, Measurement Pack, Feasibility Scorecard
CTO / Enterprise Architect	Solution Architecture Pack, Requirements Traceability Matrix
Product Leaders	PRD / Solution Brief, User Stories + Acceptance Criteria, Wireframes / Storyboards
Business Analysts	Stakeholder Register + Influence Matrix, Current-State Pack
Ops Leads	Delivery Plan (WBS + roadmap + RACI), Evidence Register
Implementation Teams	All 18 artifacts, full traceability and audit trail

Scope

MVP supports 4 connectors only:

Connector	Purpose
Notion	Docs, product specs, meeting notes, wikis
Slack	Conversations, decisions, async discussions
Jira / Linear	Issues, epics, tickets, sprint data
Salesforce	CRM data, opportunities, account hierarchies

All other systems are future adapters behind the same connector interface.

What V6 is not

No freeform document generation
No autonomous side effects
No generic chatbot
No broad connector sprawl
No production deployment automation

Canonical Data Model

V6 defines 22 canonical entity types that form the Enterprise Knowledge Model:

#	Entity	Description
1	Evidence	Raw fact with source, timestamp, confidence, provenance
2	Source	Origin system or channel (connector, upload, chat, etc.)
3	Stakeholder	Person or role with influence and interests
4	OrgUnit	Organizational unit, team, or department
5	Capability	Business capability or function
6	Process	Business process or workflow
7	System	Software system, tool, or platform
8	Requirement	Functional or non-functional requirement
9	BusinessRule	Policy, invariant, or business logic
10	Assumption	Accepted premise without full evidence
11	Constraint	Limitation or boundary condition
12	Risk	Identified risk with impact and likelihood
13	Decision	Recorded decision with rationale and date
14	Option	Considered alternative or choice
15	TradeOff	Comparison of two or more options
16	KPI	Key performance indicator with baseline/target
17	Project	Initiative, program, or workstream
18	Artifact	Generated document or deliverable
19	ArtifactVersion	Versioned snapshot of an artifact
20	Conflict	Detected contradiction between evidence/entities
21	ValidationResult	Result of a validation check
22	FeasibilityScore	Multi-dimensional feasibility score
23	AuditEvent	Immutable audit log entry

Knowledge Pipeline

Ingest → Parse → Normalize → Dedupe → Alias Resolve →
Entity/Relationship Build → Conflict Detection →
Knowledge Graph Update → Gap Analysis → Decision Analysis →
Feasibility Scoring → Artifact Projection → Validation → Publish

Each stage is a deterministic activity (Temporal) that reads from and writes to the canonical model. No stage bypasses validation.

Stage	Responsibility
Ingest	Accept raw data from 4 MVP connectors + uploads + chat
Parse	Convert raw data into typed Evidence objects
Normalize	Standardize formats, dates, references
Dedupe	Merge duplicate evidence from overlapping sources
Alias Resolve	Map aliases (e.g. "Acme Inc" ↔ "Acme Corporation")
Entity/Relationship Build	Construct canonical entities and links
Conflict Detection	Identify contradictions across sources
Knowledge Graph Update	Persist to the Enterprise Knowledge Model
Gap Analysis	Identify missing evidence or relationships
Decision Analysis	Surface decisions, options, and trade-offs
Feasibility Scoring	Score across 8 dimensions
Artifact Projection	Generate external and internal artifacts
Validation	Run all validation gates before publishing
Publish	Release artifacts to Decision Workspaces

Agent Roster

#	Agent	Responsibility
1	ChiefOfStaff	Front-door orchestrator, context assembly, routing, state merge
2	IngestionParser	Parse raw input into typed Evidence; handle all 4 connectors
3	KnowledgeMapper	Build entities, relationships, and the knowledge graph
4	ConflictResolver	Detect and resolve contradictions across sources
5	FeasibilityAnalyst	Score recommendations across 8 dimensions
6	DecisionAnalyst	Surface decisions, options, trade-offs, and rationale
7	ArtifactComposer	Template-driven artifact generation from canonical model
8	GovernanceAgent	Enforce guardrails, validation gates, and approval flows
9	EvaluationAgent	Measure quality, coverage, freshness, and calibration

Agent Rules

One responsibility per agent
Typed I/O only — no unstructured data passing between agents
No agent mutates canonical state directly (writes go through governance)
No agent bypasses validation
No agent executes side effects without governance

Artifact Catalog

External Decision Artifacts (12)

#	Artifact	Consumer
1	Executive Decision Brief	CEO, COO, Board
2	Stakeholder Register + Influence Matrix	Business Analysts, PMs
3	Org / Ownership Map	COO, Enterprise Architects
4	Current-State Pack (capability map + SIPOC + context diagram + process map)	All stakeholders
5	Feasibility & Options Report	CFO, CTO, Product Leaders
6	PRD / Solution Brief	Product, Engineering
7	User Stories + Acceptance Criteria	Engineering, QA
8	Wireframes / Storyboards	Design, Product
9	Solution Architecture Pack	CTO, Architects, Engineering
10	Requirements Traceability Matrix	QA, Compliance, PMs
11	Delivery Plan (WBS + roadmap + RACI)	Ops Leads, Implementation Teams
12	Measurement Pack (KPIs + baseline + target + retrospective)	CFO, COO

Internal Control Artifacts (6)

#	Artifact	Purpose
13	Evidence Register	All evidence with provenance, confidence, and source
14	Data Quality / Mismatch Report	Detected inconsistencies and gaps
15	Conflict Register	All unresolved and resolved conflicts
16	Decision Log	Immutable record of every decision and rationale
17	Feasibility Scorecard	Multi-dimensional scores for every recommendation
18	Artifact Health Report	Freshness, completeness, and validation status of all artifacts

Artifact Generation Rules

Template-driven — every artifact uses a deterministic template
Evidence-linked — every claim maps back to underlying Evidence
Versioned — every regeneration creates a new ArtifactVersion
Regenerated only from the canonical model — never from freeform LLM output
LLM may fill fields but may not invent sections or authoritative facts

Feasibility & Decision Readiness

Feasibility Engine

Every recommendation is scored across 8 dimensions:

Dimension	Description
Business Value	Strategic impact and ROI
Technical Complexity	Implementation difficulty and unknowns
Operational Fit	Alignment with existing processes
Financial Impact	Cost, savings, and payback period
Compliance Risk	Regulatory and policy exposure
Data Readiness	Availability and quality of required data
Change Effort	Organizational change management load
Stakeholder Alignment	Buy-in and support from affected parties

Decision Readiness Gates

Score	Action
85–100	Publish — ready for decision
70–84	Publish with caveats
60–69	Human review required
Below 60	Block and return to discovery

Validation Engine

The system blocks publication when any of the following is true:

Any requirement lacks evidence
Any claim lacks traceability
Any artifact section conflicts with the canonical model
Any KPI lacks baseline or owner
Any numeric claim fails source reconciliation
Any policy or governance rule is violated

Architecture

flowchart TD
    U["Enterprise Stakeholder"] --> WS["Decision Workspace (HTMX + SSE)"]
    WS --> CORE["Go Core (Fiber HTTP)"]

    subgraph CONNECTORS["MVP Connectors (4)"]
        N["Notion"]
        S["Slack"]
        JL["Jira / Linear"]
        SF["Salesforce"]
    end

    CONNECTORS --> ING["Ingestion Service"]
    UPL["Upload / Chat / Email / Screenshot"] --> ING
    ING --> EV["Evidence Store"]

    EV --> PIPELINE["Knowledge Pipeline (Temporal Orchestration)"]

    PIPELINE --> PARSE["Parse"]
    PIPELINE --> NORM["Normalize"]
    PIPELINE --> DEDUPE["Dedupe"]
    PIPELINE --> ALIAS["Alias Resolve"]
    PIPELINE --> ER["Entity / Relationship Build"]
    PIPELINE --> CD["Conflict Detection"]
    PIPELINE --> KG["Knowledge Graph Update"]
    PIPELINE --> GA["Gap Analysis"]
    PIPELINE --> DA["Decision Analysis"]
    PIPELINE --> FS["Feasibility Scoring"]
    PIPELINE --> AP["Artifact Projection"]
    PIPELINE --> VAL["Validation"]
    PIPELINE --> PUB["Publish"]

    KG --> EKM[("Enterprise Knowledge Model<br/>PostgreSQL + Qdrant + Graphiti")]

    AP --> ARTIFACTS["Artifact Generator"]
    ARTIFACTS --> EXT["12 External Artifacts<br/>Executive Brief, PRD, RTM, ..."]
    ARTIFACTS --> INT["6 Internal Artifacts<br/>Evidence Register, Conflict Log, ..."]

    VAL --> GATES["Decision Readiness Gates<br/>85–100 Publish | 70–84 Caveats | 60–69 Review | <60 Block"]

    PUB --> WORKSPACES["Stakeholder Decision Workspaces"]
    WORKSPACES --> CEO["CEO Dashboard"]
    WORKSPACES --> CFO["CFO Dashboard"]
    WORKSPACES --> CTO["CTO Dashboard"]
    WORKSPACES --> PM["Product Dashboard"]
    WORKSPACES --> OPS["Ops Dashboard"]

    subgraph AGENTS["Agent Roster (9)"]
        COS["ChiefOfStaff"]
        IP["IngestionParser"]
        KM["KnowledgeMapper"]
        CR["ConflictResolver"]
        FA["FeasibilityAnalyst"]
        DANA["DecisionAnalyst"]
        AC["ArtifactComposer"]
        GOV["GovernanceAgent"]
        EA["EvaluationAgent"]
    end

    CORE --> COS
    COS --> PIPELINE
    COS --> AGENTS

    subgraph TEMPORAL["Temporal Orchestration Layer"]
        WF1["IngestionWorkflow"]
        WF2["KnowledgeWorkflow"]
        WF3["FeasibilityWorkflow"]
        WF4["ArtifactWorkflow"]
        WF5["GovernanceWorkflow"]
    end

    PIPELINE --> TEMPORAL

Key components

Layer	Responsibility
Interface (`apps/core/` HTMX)	Decision Workspaces, stakeholder-specific dashboards, SSE streaming
Go Core (`apps/core/`)	Fiber HTTP server, SSE, workspace routing, Temporal client
Ingestion	4 MVP connectors (Notion, Slack, Jira/Linear, Salesforce) + upload/chat intake
Knowledge Pipeline	14-stage Temporal-orchestrated pipeline from ingest to publish
Enterprise Knowledge Model	22 canonical entity types in PostgreSQL, Qdrant, Graphiti
Agent Roster	9 agents with typed I/O, governed writes, no direct state mutation
Feasibility Engine	8-dimension scoring with Decision Readiness Gates
Validation Engine	Evidence-gated, traceability-checked, conflict-blocked publication
Artifact Projection	18 artifact templates (12 external + 6 internal)
Decision Workspaces	Stakeholder-specific views of the same canonical model
Observability	Langfuse tracing, audit events, evaluation metrics

Project Structure

apps/
  core/                         # Go Modular Monolith (control / gateway)
    cmd/
      server/                   # HTTP server entrypoint
      worker/                   # Temporal worker entrypoint
    internal/
      web/                      # HTTP handlers (Fiber + HTMX + SSE)
        handler.go              # All endpoints, workspace routing
        sse.go                  # SSE handler
        templates/
          workspaces/           # Decision Workspace views
            ceo.html
            cfo.html
            cto.html
            product.html
            ops.html
          partials/             # HTMX partials
      agents/                   # Go agent stubs
      config/                   # LLM configuration
      db/                       # sqlc generated code
      database/                 # Connection utilities
      temporal/                 # Temporal client
      workflow/                 # Temporal workflow stubs
  ai/                           # Python AI Worker
    src/
      connectors/               # V6 — MVP connector implementations
        base.py                 # Connector ABC
        notion.py
        slack.py
        jira_linear.py
        salesforce.py
      evidence/                 # V6 — Evidence ingestion and parsing
        parser.py               # Input parser (chat, upload, transcript, etc.)
        evidence_store.py       # Evidence persistence
        normalizer.py           # Normalize and dedupe
      ontology/                 # V6 — Canonical data model
        entities.py             # 22 entity types
        relationships.py        # Entity relationship definitions
        knowledge_graph.py      # Graph operations
        adapter.py              # Read/write bridge
      pipeline/                 # V6 — Knowledge Pipeline stages
        ingest.py
        parse.py
        normalize.py
        dedupe.py
        alias_resolve.py
        entity_build.py
        conflict_detection.py
        graph_update.py
        gap_analysis.py
        decision_analysis.py
        feasibility_scoring.py
        artifact_projection.py
        validation.py
        publish.py
      artifacts/                # V6 — Artifact generation
        templates/              # 18 artifact templates
        composer.py             # Template-driven artifact generation
        external/               # 12 external artifacts
        internal/               # 6 internal artifacts
      feasibility/              # V6 — Feasibility engine
        scorer.py               # 8-dimension scoring
        gates.py                # Decision Readiness Gates
      decision/                 # V6 — Decision analysis
        workspace.py            # Decision Workspace contract
        stakeholder_view.py     # Stakeholder-specific projections
      validation/               # V6 — Validation engine
        rules.py                # Validation rules
        gates.py                # Publication gates
      agents/                   # V6 — Agent roster
        chief_of_staff.py
        ingestion_parser.py
        knowledge_mapper.py
        conflict_resolver.py
        feasibility_analyst.py
        decision_analyst.py
        artifact_composer.py
        governance_agent.py
        evaluation_agent.py
      governance/               # V6 — Governance agent
        guardrails.py           # Guardrail enforcement
        approval.py             # HITL approval flows
        audit.py                # Audit event logging
      workflows/                # Temporal workflow definitions
        ingestion_workflow.py
        knowledge_workflow.py
        feasibility_workflow.py
        artifact_workflow.py
        governance_workflow.py
      mission/                  # V4.2+ — mission state, decision engine, timeline
        storage.py              # MissionStateStore (asyncpg, Go-mirrored DDL)
        timeline.py             # Frozen MissionEventType + MissionTimeline + MissionSnapshotStore (diff)
        decision.py             # MissionDecisionEngine (real LLM via src/config/llm.py)
        runtime.py              # MissionRuntime + KPI writeback
      observability/            # Hardening — typed E2E trace/correlation propagation
        trace_context.py        # TraceContext (trace_id / correlation_id / tenant_id), contextvars
        pipeline.py             # PipelineRun + run_stage (retry, latency, status)
      workspace/                # Mission / Decision workspace schema
      policy/                   # Governed policy runtime
      goal/                     # Goal runtime + models
      evaluation/               # Evaluation runtime
      session/                  # Engagement and session-state
      memory/                   # Knowledge graph, Qdrant, spine
      services/                 # Evidence layer, trust battery
      worker.py                 # Temporal worker registration
    tests/
      integration/              # Live infra suites (skip when DB/LLM absent)
        test_mission_live_infra.py       # Postgres + real LLM (8 passing)
        hardening/                       # E2E observability (live Postgres + Groq)
          test_pipeline_e2e.py
      test_mission_timeline.py   # Deterministic timeline/diff unit tests
scripts/
  hardening/                    # Chaos smoke + resource-limit documentation
    chaos_smoke.sh              # Restart postgres+temporal, assert recovery
    README.md

Getting Started

Prerequisites

Docker (Temporal, Qdrant, PostgreSQL, Neo4j/Graphiti)
Go 1.24
Python 3.13 with uv

Quickstart

# 1. Start infrastructure
make up

# 2. Run the Go server
cd apps/core && go run cmd/server/main.go

# 3. Run the Go Temporal worker
cd apps/core && go run cmd/worker/main.go

# 4. Install Python deps and run the Python worker
cd apps/ai && uv sync
cd apps/ai && uv run python -m src.worker

# 5. Open a Decision Workspace
#    http://localhost:8080

Run the tests

# Go — all packages
cd apps/core && go test ./...

# Python — full suite
cd apps/ai && uv run pytest tests/ -v

# Python — specific V6 suites
cd apps/ai && uv run pytest tests/test_pipeline/ -v
cd apps/ai && uv run pytest tests/test_connectors/ -v
cd apps/ai && uv run pytest tests/test_feasibility/ -v
cd apps/ai && uv run pytest tests/test_validation/ -v

# Python — hardening + live-infra suites (need Postgres up + GROQ_API_KEY)
cd apps/ai && set -a && . ../../.env && set +a && \
  uv run pytest tests/integration/hardening/test_pipeline_e2e.py -q -p no:cacheprovider
cd apps/ai && uv run pytest tests/integration/test_mission_live_infra.py -q -p no:cacheprovider

Environment variables

Provider	Variables
Azure AI Foundry	`AZURE_OPENAI_ENDPOINT`, `AZURE_OPENAI_API_KEY`
Groq	`GROQ_API_KEY`
OpenAI	`OPENAI_API_KEY`
Ollama (local)	`OLLAMA_BASE_URL`, `OLLAMA_API_KEY`

Control	Variable	Default
Temporal task queue	`ONTOLOGYAI-MAIN-QUEUE`	`TRACKGUARD-MAIN-QUEUE`
Live Postgres DSN	`ITERATESWARM_DATABASE_URL`	`postgresql://iterateswarm:iterateswarm@localhost:5433/iterateswarm`

Secrets live in a local .env file (never committed).

Key Metrics

V6 measures success across 12 evaluation dimensions:

Metric	Description
Parser Accuracy	% of raw input correctly parsed into typed Evidence
Entity Merge Accuracy	% of entity merges that are correct (precision/recall)
Conflict Detection Rate	% of true conflicts detected
Evidence Coverage	% of entities with ≥1 evidence link
Traceability Coverage	% of claims with traceable evidence chain
Artifact Completeness	% of required sections populated in generated artifacts
Decision Readiness	% of workspaces meeting the Decision Workspace Contract
Feasibility Calibration	Correlation between feasibility scores and actual outcomes
Human Override Rate	% of gated decisions that humans override
Stakeholder Usefulness	NPS or satisfaction score from workspace consumers
Regeneration Correctness	% of regenerated artifacts that match prior versions
Freshness	Age of evidence underpinning published artifacts

Production Hardening (Sprints A–L)

Operating rule: No new business features until every stage is observable, typed, replayable, testable, recoverable. The hardening sprints add operational qualities without new capabilities.

Sprint	Delivered
A — E2E Observability	`src/observability/` — typed `TraceContext` (trace_id/correlation_id/tenant_id) + `PipelineRun`/`run_stage` (retry, latency, status). `tests/integration/hardening/test_pipeline_e2e.py` proves one event carries the ids through connector-event → evidence → mission → decision (real Groq) → snapshot → timeline, with idempotent de-duplication on re-delivery (2/2 passing against live Postgres + real LLM).
Timeline & Versioning	`src/mission/timeline.py` — frozen `MissionEventType` verbs, append-only per-entity `timeline_events` (`UNIQUE (entity_type, entity_id, version)`), `mission_snapshots` (`UNIQUE (mission_id, version)`), and `MissionSnapshotStore.diff` (yesterday-vs-today, `confidence_delta`).
Mission Decisioning	`src/mission/decision.py` — `MissionDecisionEngine` calls only through `src/config/llm.py` (`json_mode=True`), defensive JSON parsing, safe confidence clamping.
G — Lightweight Metrics	VictoriaMetrics (single binary, Prometheus-compatible) added to `docker-compose.yml`. No Prometheus / Grafana.
J — Resource Limits	Hard `deploy.resources.limits` on every service: neo4j 4G/2cpu, redpanda 2G/1cpu, temporal 1G/1cpu, postgres 512M/1cpu, qdrant 512M/1cpu, victoriametrics 256M/0.5cpu.
Chaos Smoke	`scripts/hardening/chaos_smoke.sh` — restarts PostgreSQL + Temporal, waits up to 90s for `healthy`, exits non-zero on failure.

Infra discipline (dev): one container at a time — stop all others before testing a specific service; test, then stop. All containers use tight CPU/memory limits.

V6 Summary

Dimension	V6
Vision	Decision Intelligence Platform
Connectors	4 MVP (Notion, Slack, Jira/Linear, Salesforce)
Canonical Entities	22 (Evidence, Stakeholder, Decision, KPI, etc.)
Pipeline Stages	14 (Ingest → Parse → ... → Publish)
Agents	9 (ChiefOfStaff, IngestionParser, ..., EvaluationAgent)
External Artifacts	12 (Executive Brief, PRD, RTM, Delivery Plan, etc.)
Internal Artifacts	6 (Evidence Register, Conflict Log, etc.)
Feasibility Dimensions	8 (Business Value, Technical Complexity, etc.)
Decision Gates	4 tiers (85–100 publish, <60 block)
Orchestration	Temporal (replay-safe, deterministic, continue-as-new)
Guardrails	No evidence = no artifact; unresolved conflict = reject

Test Coverage

Suite	Scope
Unit Tests	Parsers, merge logic, scoring, template rendering, validators
Integration Tests	Connectors, evidence flow, artifact projection, approval gates
Temporal Tests	Replay, pause/resume, retries, signals, updates, continue-as-new
Adversarial Tests	Conflicting sources, missing fields, duplicates, stale data, prompt injection, malformed files, inconsistent numbers
End-to-End Tests	Executive Brief → PRD → Architecture → RTM → Delivery Plan → Measurement Pack
Hardening Tests	E2E observability (live Postgres + Groq), mission timeline/diff unit suites, Postgres live-infra suite (8 passing)
Go Build	✅ Clean
Python Unit Tests	✅ Baseline + V6 suites

Architecture Decision Records

Key ADRs governing V6 design:

ADR-001: Evidence-first pipeline — no artifact without evidence chain
ADR-002: Temporal as single orchestration backbone — activities are deterministic and replayable
ADR-003: Governance exclusivity — only GovernanceAgent finalizes side effects
ADR-004: Template-driven artifacts — LLM fills fields, never invents sections
ADR-005: Decision Readiness Gates — publication is gated by scored readiness
ADR-006: Connector interface — abstract adapter pattern for all future connectors beyond the MVP 4
ADR-007: Context engineering — layered context assembly never passes raw unbounded history
ADR-008: 22-entity canonical model — single source of truth for all projections

Contributing & Conventions

Feature branches: git checkout -b feature/description — never commit directly to main.
Conventional Commits: feat:, fix:, refactor:, docs:, test:, chore:.
Never commit to main. Open a PR from your feature branch.
Secrets: use a local .env file; never commit secrets.
Agent guidelines: full coding standards and build/test commands in AGENTS.md.
Definition of Done: A feature is done only when it updates the canonical model, survives validation, appears correctly in all affected workspaces, passes tests, carries traceability, and can be replayed from evidence.

Name		Name	Last commit message	Last commit date
Latest commit History 242 Commits
.githooks		.githooks
.github/workflows		.github/workflows
.opencode		.opencode
apps		apps
archive		archive
config		config
docker		docker
docs		docs
k8s/helm/sarthi		k8s/helm/sarthi
migrations		migrations
mockoon		mockoon
playwright		playwright
proto/ai/v1		proto/ai/v1
scripts		scripts
tests/mockoon		tests/mockoon
.env.example		.env.example
.gitignore		.gitignore
AGENTS.md		AGENTS.md
AGENT_INSTRUCTION.md		AGENT_INSTRUCTION.md
ARCHITECTURE.md		ARCHITECTURE.md
CAPABILITY_ONTOLOGY_V6.md		CAPABILITY_ONTOLOGY_V6.md
CODEBASE_EXPLANATION.md		CODEBASE_EXPLANATION.md
CODEBASE_WALKTHROUGH.md		CODEBASE_WALKTHROUGH.md
IMPLEMENTATION_PLAN.md		IMPLEMENTATION_PLAN.md
Makefile		Makefile
PLAN_V5.1.md		PLAN_V5.1.md
PLAN_V6.md		PLAN_V6.md
PRODUCTION_TEST_CHECKLIST.md		PRODUCTION_TEST_CHECKLIST.md
Procfile		Procfile
README.md		README.md
REPO_MAPPING.md		REPO_MAPPING.md
RISK_REGISTER.md		RISK_REGISTER.md
V6_DOMAIN_SPEC.md		V6_DOMAIN_SPEC.md
audit_plan.md		audit_plan.md
buf.gen.yaml		buf.gen.yaml
buf.yaml		buf.yaml
build.sh		build.sh
docker-compose.prod.yml		docker-compose.prod.yml
docker-compose.startup-guardian.yml		docker-compose.startup-guardian.yml
docker-compose.test.yml		docker-compose.test.yml
docker-compose.v3.yml		docker-compose.v3.yml
docker-compose.yml		docker-compose.yml
e2e_chat_test.mjs		e2e_chat_test.mjs
fix_mockoon_routes.py		fix_mockoon_routes.py
mockoon.json		mockoon.json
prd.md		prd.md
pytest.ini		pytest.ini
render.yaml		render.yaml
repo2pdf.ignore		repo2pdf.ignore
sarthi-command-center.html		sarthi-command-center.html

Navigation Menu

Repository files navigation

OntologyAI V6 — Decision Intelligence Platform

Table of Contents

Vision

Product Principle

Who it is for

Scope

What V6 is not

Canonical Data Model

Knowledge Pipeline

Agent Roster

Agent Rules

Artifact Catalog

External Decision Artifacts (12)

Internal Control Artifacts (6)

Artifact Generation Rules

Feasibility & Decision Readiness

Feasibility Engine

Decision Readiness Gates

Validation Engine

Architecture

Key components

Project Structure

Getting Started

Prerequisites

Quickstart

Run the tests

Environment variables

Key Metrics

Production Hardening (Sprints A–L)

V6 Summary

Test Coverage

Architecture Decision Records

Contributing & Conventions

About

Releases

Packages

Contributors

Languages

Folders and files

Latest commit

History

Repository files navigation

OntologyAI V6 — Decision Intelligence Platform

Table of Contents

Vision

Product Principle

Who it is for

Scope

What V6 is not

Canonical Data Model

Knowledge Pipeline

Agent Roster

Agent Rules

Artifact Catalog

External Decision Artifacts (12)

Internal Control Artifacts (6)

Artifact Generation Rules

Feasibility & Decision Readiness

Feasibility Engine

Decision Readiness Gates

Validation Engine

Architecture

Key components

Project Structure

Getting Started

Prerequisites

Quickstart

Run the tests

Environment variables

Key Metrics

Production Hardening (Sprints A–L)

V6 Summary

Test Coverage

Architecture Decision Records

Contributing & Conventions

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages