┌─────────────────────────────────────────────────────────────────┐
│                                                                 │
│      █████╗  ██████╗ ███████╗███╗   ██╗████████╗██╗ ██████╗    │
│     ██╔══██╗██╔════╝ ██╔════╝████╗  ██║╚══██╔══╝██║██╔════╝    │
│     ███████║██║  ███╗█████╗  ██╔██╗ ██║   ██║   ██║██║         │
│     ██╔══██║██║   ██║██╔══╝  ██║╚██╗██║   ██║   ██║██║         │
│     ██║  ██║╚██████╔╝███████╗██║ ╚████║   ██║   ██║╚██████╗   │
│     ╚═╝  ╚═╝ ╚═════╝ ╚══════╝╚═╝  ╚═══╝   ╚═╝   ╚═╝ ╚═════╝   │
│                                                                 │
│          Enterprise AI Agent Orchestration Platform              │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

Overview

Agentics CLI is an enterprise-grade AI agent orchestration platform that enables organizations to plan, simulate, deploy, and govern AI agent infrastructure at scale. It provides a unified command-line interface for managing the full lifecycle of AI agent deployments -- from initial simulation planning through compliance packaging and infrastructure export.

The platform orchestrates 27 deployed agent domains comprising 109 specialized agents that cover security, cost optimization, observability, governance, ERP integration, and more. Agents are dispatched in parallel through 33 auto-dispatched multi-agent combinations, enabling complex enterprise workflows to be resolved through a single natural language command.

Agentics CLI integrates directly with enterprise ERP systems including SAP S/4HANA, Oracle ERP Cloud, and Microsoft Dynamics 365, and exports production-ready infrastructure-as-code for Terraform, Kubernetes, CloudFormation, Pulumi, and Ansible. All operations produce auditable artifacts suitable for SOC2, HIPAA, GDPR, and PCI-DSS compliance reviews.

Key Features

6-Phase Enterprise Pipeline -- Structured workflow from simulation through deployment with full traceability
27 Agent Domains, 109 Agents -- Specialized agents for security, cost analysis, observability, governance, and more
Multi-Agent Orchestration -- 33 pre-configured agent combinations dispatched in parallel via agentics ask
Natural Language Interface -- Describe tasks in plain English; the CLI classifies intent and routes automatically
10 ERP System Integrations -- Native connectors for SAP, Oracle, NetSuite, Dynamics 365, and more
Infrastructure-as-Code Export -- Generate Terraform, Kubernetes, CloudFormation, Pulumi, and Ansible artifacts
Compliance Packaging -- Automated SOC2, HIPAA, GDPR, and PCI-DSS artifact generation
CFO-Grade Financial Analysis -- ROI quantification with executive-summary and detailed reporting
MCP Server Integration -- Operates as a Model Context Protocol server for AI-native toolchains
Deterministic Simulation -- Reproducible simulation runs with configurable iteration counts and parallelism

Architecture

6-Phase Enterprise Pipeline

┌───────────┐    ┌───────────┐    ┌───────────┐    ┌───────────┐    ┌───────────┐    ┌───────────┐
│           │    │           │    │           │    │           │    │           │    │           │
│   PLAN    │───>│ SIMULATE  │───>│  INSPECT  │───>│ QUANTIFY  │───>│  DEPLOY   │───>│  EXPORT   │
│           │    │           │    │           │    │           │    │           │    │           │
└───────────┘    └───────────┘    └───────────┘    └───────────┘    └───────────┘    └───────────┘
      │                │                │                │                │                │
      v                v                v                v                v                v
  Simulation       Iteration        Metrics &        Financial       Deployment       IaC Artifacts
   Planner          Runner           Outputs        ROI Engine         Intent         & Compliance

Agent Routing

  agentics ask "<query>"
        │
        ├── Combo Dispatch (33 multi-agent sets, parallel)
        ├── Single Agent Match (95 keyword rules)
        └── Direct Invoke (agentics agents invoke <domain> <agent>)
        │
        v
  27 Agent Domains ── 109 Specialized Agents

Quick Start

Installation

npm install -g @llm-dev-ops/agentics-cli

Or run directly without installation:

npx @llm-dev-ops/agentics-cli help

Authentication

agentics login

First Command

agentics ask "What is the cost breakdown for our AI agent infrastructure?"

Note: All workflow commands accept natural language input in quotes. Structured JSON references are also supported for programmatic use.

Core Workflow

The CLI follows a structured 6-phase pipeline for AI agent deployment. Each phase produces artifacts that feed into the next.

Phase	Command	Description	Upstream Service
1	`agentics plan`	Create a simulation plan from an organization manifest	agentics-simulation-planner
2	`agentics simulate`	Execute simulation iterations against the plan	agentics-simulation-runner
3	`agentics inspect`	Retrieve deterministic outputs and metrics	agentics-simulation-engine
4	`agentics quantify`	Generate CFO-grade ROI and financial analysis	enterprise-roi-engine
5	`agentics deploy`	Resolve deployment intent from simulation results	agentics-deployment-intent
6	`agentics export`	Generate infrastructure-as-code artifacts	agentics-deployment-exporters

An additional diligence command packages compliance artifacts across all phases.

Example: End-to-End Workflow

# 1. Create a deployment plan
agentics plan "prod-cluster-v2" --params '{"region":"us-east-1"}'

# 2. Run 10 simulation iterations
agentics simulate '{"id":"plan-abc123"}' --iterations 10 --config '{"parallelism":4}'

# 3. Retrieve simulation metrics
agentics inspect '{"id":"sim-xyz789"}' --output-type metrics --format table

# 4. Generate financial analysis
agentics quantify '{"id":"sim-xyz789"}' --report-type cfo-grade \
  --params '{"currency":"USD","timeframe":"annual"}'

# 5. Deploy to staging
agentics deploy '{"id":"sim-xyz789"}' --environment staging

# 6. Export Terraform configurations
agentics export '{"id":"intent-456"}' --export-format terraform --output ./infrastructure/

# 7. Package compliance documentation
agentics diligence '[{"id":"sim-xyz789"},{"id":"intent-456"}]' --frameworks SOC2,HIPAA

Commands Reference

Workflow Commands

`agentics plan`

Create a simulation plan from an organization manifest.

agentics plan <manifest-query> [--params <json>]

`agentics simulate`

Execute a simulation from a plan reference or natural language description.

agentics simulate <plan-ref|description> [--config <json>] [--iterations <n>]

`agentics inspect`

Retrieve deterministic outputs from a simulation.

agentics inspect <sim-ref|description> [--output-type <type>]

Output types: metrics, logs, state, events, summary, full

`agentics quantify`

Generate financial impact and ROI analysis.

agentics quantify <sim-ref|description> [--report-type <type>] [--params <json>]

Report types: executive-summary, detailed-analysis, cfo-grade, custom

`agentics deploy`

Resolve deployment intent from simulation results.

agentics deploy <sim-ref|description> [--spec <json>] [--environment <env>]

`agentics export`

Generate infrastructure-as-code deployment artifacts.

agentics export <intent-ref|description> [--export-format <format>] [--output <path>]

Export formats: terraform, kubernetes, cloudformation, pulumi, ansible, custom

`agentics diligence`

Package compliance and audit artifacts.

agentics diligence <refs|description> [--frameworks <list>] [--requirements <json>]

Frameworks: SOC2, HIPAA, GDPR, PCI-DSS, custom

Agent Commands

`agentics ask`

Route a natural language query to the appropriate agent(s). Automatically dispatches multi-agent combinations when applicable.

agentics ask "<natural language query>"

`agentics agents list`

List all available agents across all domains.

agentics agents list

`agentics agents invoke`

Invoke a specific agent directly by domain and agent name.

agentics agents invoke <domain> <agent> '{"text":"<prompt>"}'

ERP Integration Commands

Command	Description
`agentics erp list`	List all supported ERP systems and connection status
`agentics erp surface <system>`	Surface available modules and data structures
`agentics erp map <system> [--modules <list>]`	Map ERP fields to the Agentics data model
`agentics erp inspect <system>`	Inspect connection details and data availability
`agentics erp export <system> [--format <fmt>]`	Export data mappings for infrastructure generation

Governance Commands

Command	Description
`agentics policy create <name> [--rules <json>]`	Create a governance policy
`agentics policy list`	List all policies
`agentics policy inspect <id>`	View policy details
`agentics policy enable <id>`	Enable a policy
`agentics policy disable <id>`	Disable a policy
`agentics policy delete <id>`	Delete a policy
`agentics audit [--from <date>] [--to <date>]`	Generate audit trails

Observability Commands

Command	Description
`agentics inspect run <run-id>`	Inspect a specific simulation or deployment run
`agentics inspect artifacts <run-id>`	List artifacts produced by a run
`agentics inspect query [--status <s>]`	Query across runs with filters
`agentics usage [--period <period>]`	View usage statistics and quota information

Account Commands

Command	Description
`agentics login`	Authenticate with the Agentics platform
`agentics logout`	Clear stored credentials
`agentics whoami`	Display current authenticated identity
`agentics version`	Display CLI and platform version

Global Options

Option	Description
`--timeout <ms>`	Request timeout in milliseconds
`--trace-id <id>`	Correlation ID for distributed tracing
`--format <fmt>`	Output format: `json`, `yaml`, `table`, `text`, `csv`
`--verbose, -v`	Enable verbose output
`--pretty`	Pretty-print output
`--version`	Show CLI version
`--help`	Display command help

Multi-Agent Orchestration

The Agentics platform organizes 109 specialized agents across 27 domains. The agentics ask command automatically classifies user intent and dispatches the optimal set of agents in parallel.

Agent Domains

Domain	Agent Count	Capability
shield	9	Security scanning, PII detection, content safety, credential exposure
test-bench	14	Benchmarking, regression testing, red-teaming, bias and hallucination detection
costops	5	Cost attribution, forecasting, budgeting, ROI analysis
observatory	7	Telemetry, usage patterns, health checks, SLO monitoring
orchestrator	7	Workflow scheduling, dependency management, parallel execution
sentinel	5	Anomaly detection, drift monitoring, root cause analysis
copilot	7	Task planning, intent classification, decomposition
memory-graph	6	Conversation history, decision lineage, knowledge graphs
connector-hub	5	ERP integration, database queries, webhooks, SSO/OAuth
auto-optimizer	3	Token optimization, model selection, self-optimization
edge	5	Circuit breakers, failover, execution guards, caching
forge	4	SDK generation, CLI tooling, API translation
platform	4	Executive summaries, decision memos, risk scoring
governance-dashboard	3	Audit trails, impact analysis, oversight
policy-engine	3	Policy enforcement, constraints, approval workflows
incident-manager	3	Escalation, post-mortems, human-in-the-loop
data-vault	2	Access control, data anonymization
simulator	2	What-if analysis, scenario simulation
latency-lens	2	Latency profiling, cold-start analysis
research-lab	2	Hypothesis testing, metrics analysis
analytics-hub	2	Consensus building, recommendation reports
registry	3	Agent indexing, reputation scoring, bootstrapping
marketplace	2	Agent packaging, deprecation management
config-manager	1	Configuration validation
schema-registry	1	Schema validation
benchmark-exchange	1	Benchmark publishing
inference-gateway	1	Inference routing and load balancing

Multi-Agent Combinations

When a query matches a known pattern, agentics ask dispatches multiple agents in parallel. Selected examples:

Combination	Trigger Patterns	Agents Dispatched
cost-analysis	"cost analysis", "spending report"	costops/attribution + forecast + roi
security-scan	"security scan", "safety scan"	shield/prompt-injection + pii + secrets + toxicity + credential-exposure
model-evaluation	"evaluate model", "model quality"	test-bench/benchmark + quality + hallucination + faithfulness
incident-response	"service down", "outage"	incident-manager/escalation + post-mortem + sentinel/rca + observatory/failures
executive-brief	"executive summary", "board report"	platform/executive-summary + costops/attribution + observatory/slo + research-lab/metrics
compliance-review	"compliance check", "GDPR", "HIPAA"	governance-dashboard/audit + oversight + policy-engine/enforce + data-vault/access-control
decision-package	"go/no-go", "decision package"	platform/decision + decision-memo + risk-score + costops/roi
latency-debug	"why is it slow", "latency issue"	latency-lens/latency + cold-start + sentinel/anomaly + rca
resilience-check	"handle failures", "make resilient"	edge/circuit-breaker + failover + execution-guard + orchestrator/retry
enterprise-data	"query SAP", "ERP data"	connector-hub/database-query + erp-surface + auth-identity

Note: 33 multi-agent combinations are available in total. Run agentics agents list for the complete catalog.

ERP Integration

The platform provides native connectors for 10 enterprise ERP systems through the connector-hub agent domain.

ERP System	API Type	Key Modules
SAP S/4HANA	OData / RFC	Finance, Materials Management, Sales, HR, Plant Maintenance
Oracle ERP Cloud	REST	General Ledger, Procurement, Project Management, Fixed Assets
NetSuite	SuiteTalk / REST	Financial Management, CRM, Inventory, E-commerce
Microsoft Dynamics 365	Dataverse / OData	Finance, Supply Chain, Commerce, Human Resources
Infor CloudSuite	ION API	Manufacturing, Distribution, Finance, HCM
Epicor Kinetic	REST	Manufacturing, Distribution, Financial Management
Sage X3	REST / SOAP	Finance, Distribution, Manufacturing
IFS Cloud	REST	Asset Management, Field Service, Manufacturing, Projects
Workday	REST / SOAP	HCM, Financial Management, Planning, Payroll
PeopleSoft	Component Interface	HCM, Finance, Supply Chain, Campus Solutions

# List available ERP systems
agentics erp list

# Surface modules from SAP
agentics erp surface sap-s4hana

# Map ERP fields to Agentics data model
agentics erp map dynamics-365 --modules finance,supply-chain

# Export mappings as Terraform
agentics erp export oracle-erp-cloud --format terraform

Compliance and Governance

Agentics CLI supports automated compliance artifact generation for major regulatory frameworks.

Framework	Coverage
SOC2	Trust Services Criteria (Security, Availability, Processing Integrity, Confidentiality, Privacy)
HIPAA	Administrative, Physical, and Technical Safeguards; Breach Notification
GDPR	Data Processing Records, DPIA, Data Subject Rights, Cross-Border Transfer
PCI-DSS	Cardholder Data Protection, Access Control, Network Security, Monitoring

# Generate compliance package
agentics diligence "package compliance artifacts for SOC2 and HIPAA audit"

# Run governance audit
agentics audit --from 2025-01-01 --to 2025-12-31 --format json

# Create and enforce policies
agentics policy create data-retention --rules '{"max_days":90,"pii_handling":"encrypt"}'
agentics policy enable data-retention

Important: Compliance artifacts are generated based on platform telemetry and simulation data. They should be reviewed by qualified compliance personnel before submission to auditors.

Infrastructure Export

The export command generates production-ready infrastructure-as-code from deployment intents.

Format	Output	Use Case
Terraform	`.tf` files with provider configurations	Multi-cloud infrastructure provisioning
Kubernetes	YAML manifests (Deployments, Services, ConfigMaps)	Container orchestration
CloudFormation	JSON/YAML templates	AWS-native infrastructure
Pulumi	TypeScript/Python programs	Programmatic infrastructure
Ansible	Playbooks and role definitions	Configuration management

# Export Terraform for a deployment intent
agentics export '{"id":"intent-456"}' --export-format terraform --output ./infra/terraform/

# Export Kubernetes manifests
agentics export "generate kubernetes manifests for the production cluster" --output ./infra/k8s/

# Export CloudFormation template
agentics export '{"id":"intent-456"}' --export-format cloudformation --output ./infra/cfn/

Natural Language Interface

All workflow commands accept natural language descriptions in place of structured JSON references. The CLI automatically detects input type and routes accordingly.

# Simulation planning
agentics plan "production deployment for the payments microservice"

# Simulation execution
agentics simulate "run an enterprise ERP cost optimization simulation"

# Results inspection
agentics inspect "show me the results from the last production test"

# Financial analysis
agentics quantify "calculate ROI for the Q4 infrastructure upgrade"

# Deployment
agentics deploy "deploy the microservices update to staging"

# IaC generation
agentics export "generate terraform for the new kubernetes cluster"

# Compliance packaging
agentics diligence "package compliance docs for the SOC2 audit"

# Multi-agent queries
agentics ask "what is the cost vs quality tradeoff for switching to a smaller model?"

Input Detection Rules:

Input beginning with { or [ is parsed as structured JSON
All other input is treated as a natural language description
Malformed JSON is treated as natural language

Environment Variables

Variable	Description	Default
`AGENTICS_API_KEY`	Platform API key for authentication	--
`AGENTICS_PLATFORM_URL`	Base URL for the Agentics platform	`https://platform.agentics.dev`
`AGENTICS_MANIFESTS_URL`	Organization manifests service endpoint	Platform default
`AGENTICS_PLANNER_URL`	Simulation planner service endpoint	Platform default
`AGENTICS_RUNNER_URL`	Simulation runner service endpoint	Platform default
`AGENTICS_SIMULATION_ENGINE_URL`	Simulation engine service endpoint	Platform default
`ENTERPRISE_ROI_ENGINE_URL`	ROI engine service endpoint	Platform default
`AGENTICS_INTENT_URL`	Deployment intent service endpoint	Platform default
`AGENTICS_EXPORTERS_URL`	Deployment exporters service endpoint	Platform default
`DILIGENCE_URL`	Diligence artifacts service endpoint	Platform default
`ANTHROPIC_API_KEY`	Anthropic API key (for NL classification)	--
`OPENAI_API_KEY`	OpenAI API key (alternative NL provider)	--

Exit Codes

Code	Name	Description
`0`	Success	Command completed successfully
`1`	General Error	Unspecified error
`64`	Usage Error	Invalid command syntax or arguments
`65`	Data Format Error	Invalid input data format
`69`	Service Unavailable	Upstream service is unreachable
`70`	Internal Error	Unexpected internal error
`100`	Validation Error	Argument or schema validation failed
`120`	Network Error	Network connectivity failure
`130`	Authentication Error	Invalid or expired credentials
`140`	Service Error	Upstream service returned an error

API Reference

Complete API documentation is available on the Agentics platform:

Platform Console: https://platform.agentics.dev
API Documentation: https://platform.agentics.dev/docs
CLI Help: Run agentics help or agentics <command> --help

The CLI also operates as a Model Context Protocol (MCP) server, enabling integration with AI-native development environments and agent toolchains.

Security

The Agentics CLI implements the following security measures:

Authentication -- API key and token-based authentication with the platform
Transport Security -- All communications use TLS 1.2+
Credential Storage -- Credentials are stored in the user's OS keychain or local configuration with restricted file permissions
Input Validation -- All inputs are validated using Zod schemas before processing
Audit Logging -- All agent invocations and data access operations are logged for audit
Policy Enforcement -- Configurable policies govern agent execution boundaries and data access

To report security vulnerabilities, contact security@globalbusinessadvisors.com.

Contributing

Contributions are welcome. To contribute:

Fork the repository at https://github.com/llm-dev-ops/agentics-cli
Create a feature branch from main
Ensure all tests pass: npm test
Ensure type checking passes: npm run typecheck
Ensure linting passes: npm run lint
Submit a pull request with a clear description of the changes

Development Setup

git clone https://github.com/llm-dev-ops/agentics-cli.git
cd agentics-cli
npm install
npm run build
npm test

Support

Channel	Link
Platform	https://platform.agentics.dev
GitHub Issues	https://github.com/llm-dev-ops/agentics-cli/issues
Repository	https://github.com/llm-dev-ops/agentics-cli
npm	https://www.npmjs.com/package/@llm-dev-ops/agentics-cli

Requirements

Node.js >= 18.0.0
npm >= 9.0.0 (included with Node.js 18+)
Operating Systems: macOS, Linux, Windows (WSL2 recommended)

License

For licensing inquiries, contact Global Business Advisors.

Keywords

FAQs

What is @llm-dev-ops/agentics-cli?

Is @llm-dev-ops/agentics-cli well maintained?

Package last updated on 01 Apr 2026

Did you know?

Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.

Install

@llm-dev-ops/agentics-cli

Agentics CLI