Eldric 4.0.0 Module

Agent Generator

Factory for Specialized AI Workers

The Agent Generator Module is an automated factory that creates specialized AI worker agents with domain-specific capabilities, custom tool sets, and knowledge bases. Define your requirements and generate production-ready agents with a single API call.

Agent Generator Architecture

INPUT Agent Spec YAML/JSON Domain Config Tools & Perms Knowledge Base Documents/RAG Constraints Limits & Policies GENERATOR CORE Schema Validator Template Engine Tool Injector Prompt Synthesizer Code Generator BUILD SYSTEM Docker Build Multi-platform Test Runner Validation Packager DEB/RPM/PKG Signer Cryptographic OUTPUT Generated Agent .eldric-agent Config Files Systemd Units Documentation

What Is the Agent Generator?

Template-Based Generation

Define agent specifications using YAML or JSON schemas. The Generator uses intelligent templates to produce consistent, production-ready agent configurations with proper system prompts, tool bindings, and safety constraints.

Domain-Specific Tool Injection

Each generated agent receives a curated set of tools based on its domain. A financial agent gets market APIs and calculation tools, while a medical agent gets clinical databases and diagnostic utilities - automatically configured.

Knowledge Base Integration

Attach domain-specific knowledge bases during generation. The agent automatically receives RAG capabilities with pre-indexed documentation, manuals, procedures, or any corpus relevant to its specialty.

Safety Constraints Built-In

Generated agents come with configurable safety guardrails - rate limits, allowed actions, forbidden topics, and audit logging. Enterprise compliance requirements are templated and automatically applied.

Agent Generation Flow

1 SPEC Define Agent YAML/JSON spec 2 VALIDATE Schema Check Verify config 3 GENERATE Create Agent Prompt + Tools 4 INJECT Add Knowledge RAG + Context 5 PACKAGE Build & Sign Deployable .agent .conf .service .md Input: Agent specification (name, domain, tools, knowledge, constraints) Output: Production-ready agent package with config, systemd unit, documentation, and deployment scripts

Use Cases

Financial Analyst Agent

Generate agents specialized in market analysis, portfolio management, risk assessment, and regulatory compliance reporting.

Market APIs Risk Calc SEC Filings Bloomberg

Medical Research Agent

Create agents for clinical literature review, drug interaction analysis, patient data processing with HIPAA compliance built-in.

PubMed ClinicalTrials HIPAA FDA

DevOps Automation Agent

Generate agents for CI/CD pipeline management, infrastructure monitoring, incident response, and automated deployments.

Kubernetes Terraform AWS/GCP Prometheus

Legal Research Agent

Build agents for case law research, contract analysis, compliance checking, and legal document drafting with citation tracking.

Westlaw LexisNexis Contract AI Citations

Scientific Data Agent

Generate agents for bioinformatics analysis, astronomical data processing, materials science simulation, and research paper synthesis.

NCBI NASA ArXiv CERN

Customer Support Agent

Create agents specialized in product knowledge, ticket routing, sentiment analysis, and multi-language customer communication.

Zendesk Salesforce i18n Sentiment

Domain-Specific Tool Injection

Agent Generator Finance Bloomberg API Risk Calculator SEC EDGAR Portfolio Tools Medical PubMed API Drug Database HIPAA Checker Clinical Tools DevOps Kubernetes API Terraform AWS/GCP SDK Monitoring Legal Case Law DB Contract Parser Citation Tracker Compliance Science NCBI/BLAST NASA APIs ArXiv Search Simulations Support Ticket System Knowledge Base Sentiment API Translation

The 6-Phase Generation Pipeline

Each agent generation request flows through a sophisticated multi-phase pipeline that ensures consistency, quality, and production-readiness.

1

Schema Parse

Parse and validate YAML/JSON specification against agent schema

2

Template Select

Choose base template based on domain and capability requirements

3

Tool Injection

Inject domain-specific tools with proper configurations

4

Prompt Synthesis

Generate optimized system prompt with safety constraints

5

Knowledge Bind

Attach RAG knowledge bases and configure retrieval

6

Package & Sign

Create deployable package with configs and documentation

Quick Start

# Generate a new specialized agent curl -X POST http://localhost:8880/api/v1/agent-generator/generate \ -H "Content-Type: application/json" \ -d '{ "name": "financial-analyst", "domain": "finance", "description": "Expert in market analysis and portfolio optimization", "tools": ["market-data", "risk-calculator", "sec-filings"], "knowledge_bases": ["financial-regulations", "market-history"], "constraints": { "rate_limit": 100, "allowed_actions": ["read", "analyze", "report"], "compliance": ["SOX", "SEC"] } }' # Response: {"agent_id": "agt-fin-abc123", "status": "generating"} # Check generation status curl http://localhost:8880/api/v1/agent-generator/status/agt-fin-abc123 # Download generated agent package curl -O http://localhost:8880/api/v1/agent-generator/download/agt-fin-abc123 # List available domain templates curl http://localhost:8880/api/v1/agent-generator/templates # List available tool sets by domain curl http://localhost:8880/api/v1/agent-generator/tools?domain=finance

Agent Specification Schema

# Example: agent-spec.yaml name: "medical-research-assistant" version: "1.0.0" domain: "medical" description: | Specialized agent for clinical literature review and drug interaction analysis. HIPAA compliant with audit logging enabled. model: preferred: "llama3.2:70b" fallback: "qwen2.5:32b" temperature: 0.3 max_tokens: 4096 tools: - "pubmed-search" - "drug-interaction-checker" - "clinical-trials-api" - "fda-adverse-events" knowledge_bases: - id: "medical-guidelines" type: "rag" embedding_model: "nomic-embed-text" - id: "drug-database" type: "structured" constraints: rate_limit: 50 max_context_tokens: 128000 allowed_actions: - "search" - "analyze" - "summarize" - "cite" forbidden_topics: - "prescriptions" - "diagnoses" compliance: - "HIPAA" - "FDA-21-CFR-11" output: format: "package" include_systemd: true include_docs: true platforms: - "linux_deb" - "linux_rpm"

API Reference

Method Endpoint Description
POST /api/v1/agent-generator/generate Generate a new agent from specification
GET /api/v1/agent-generator/status/:id Get generation job status and progress
GET /api/v1/agent-generator/download/:id Download generated agent package
GET /api/v1/agent-generator/templates List available domain templates
GET /api/v1/agent-generator/templates/:domain Get template details for a domain
GET /api/v1/agent-generator/tools List available tools (filter by domain)
GET /api/v1/agent-generator/knowledge-bases List available knowledge bases
POST /api/v1/agent-generator/validate Validate agent specification without generating
POST /api/v1/agent-generator/preview Preview generated prompt without full build
DELETE /api/v1/agent-generator/jobs/:id Cancel a generation job
GET /api/v1/agent-generator/agents List all generated agents
GET /api/v1/agent-generator/agents/:id Get generated agent details

Agent Generator vs Agent Builder

Agent Generator

Template-based
Fast generation
Deterministic output
Predefined domains

vs

Use Generator when you have a clear domain and want fast, consistent results
Use Builder when you need custom reasoning and novel agent architectures

Agent Builder

LLM-powered design
Iterative refinement
Creative solutions
Novel architectures

License Features

Feature Free Standard Professional Enterprise
Agents per month 5 50 500 Unlimited
Domain templates 3 All All All + Custom
Tool sets Basic Standard Advanced All + Custom
Knowledge base size 100MB 1GB 10GB Unlimited
Package formats 1 All All All + Custom
Compliance templates - Basic All All + Custom
API rate limit 10/hr 100/hr 1000/hr Unlimited