What is the Logic Register?

A central registry that stores predicates, facts, and rules with metadata such as domain, type, signature, and description, enabling deduplication and cross-domain linking.

How does deduplication work?

It checks for exact name matches, analyzes semantic similarity in descriptions, suggests existing predicates, and maintains aliases to cover synonyms.

What integrations does it use?

It uses a logic-registry MCP server for storage, semantic similarity for duplicate detection, and the file system for backups and exports.

Logic Register

npx machina-cli add skill NewJerseyStyle/plugin-logic-llm/logic-register --openclaw

Files (1)

SKILL.md

2.8 KB

/logic-register

Description

This skill manages the central registry that tracks all predicates, facts, and rules across the knowledge base. It ensures naming consistency, prevents duplicates, and enables code reuse across conversions.

Usage

/logic-register [command] [options]

Commands

add - Register new predicates or rules
search - Search for existing predicates
list - List registered items
update - Update predicate descriptions or metadata
link - Link related predicates across domains
export - Export registry to file

Options

--domain <name> - Filter by domain
--type <predicate|fact|rule> - Filter by type
--format <json|prolog|clingo> - Output format

Examples

/logic-register add --domain legal "canAccess(Person, Resource) ::: Person is authorized to access Resource"
/logic-register search "access"
/logic-register list --domain legal --type predicate
/logic-register export --format json ./registry-backup.json

Registry Structure

Each registry entry contains:

{
  "name": "canAccess",
  "arity": 2,
  "domain": "legal",
  "type": "predicate",
  "signature": "canAccess(Person, Resource)",
  "description": "Person is authorized to access Resource",
  "arguments": [
    {"name": "Person", "type": "entity", "description": "The person requesting access"},
    {"name": "Resource", "type": "entity", "description": "The resource being accessed"}
  ],
  "examples": ["canAccess(john, confidential_files)"],
  "source": "contracts/access-policy.txt",
  "created": "2024-01-15T10:30:00Z",
  "aliases": ["hasAccess", "isAuthorized"],
  "related": ["hasRole", "resourceClassification"]
}

Deduplication

The registry prevents duplicate predicates by:

Checking for exact name matches
Checking for semantic similarity in descriptions
Suggesting existing predicates when creating new ones
Maintaining aliases for the same concept

Cross-Domain Linking

Predicates can be linked across domains:

/logic-register link legal.canAccess hr.hasEmploymentStatus

This enables queries that span multiple knowledge domains.

Integration

This skill uses:

logic-registry MCP server for storage
Semantic similarity for duplicate detection
File system for backup/export

Best Practices

Always search before adding new predicates
Use descriptive names following domain conventions
Include examples for complex predicates
Link related predicates for better query resolution
Regularly export backups of the registry

Notes

The registry persists across sessions in data/registry/
Registry entries include provenance (source document)
Use aliases to handle synonyms and variations
The registry supports versioning for predicate evolution

Source

git clone https://github.com/NewJerseyStyle/plugin-logic-llm/blob/main/skills/logic-register/SKILL.mdView on GitHub

Overview

Logic Register maintains the central registry that tracks all predicates, facts, and rules across the knowledge base. It enforces naming consistency, prevents duplicates, and enables code reuse across conversions.

How This Skill Works

Entries are stored in a logic-registry MCP-backed store and accessed via /logic-register commands (add, search, list, update, link, export). It uses semantic similarity to detect duplicates and supports cross-domain linking to connect related predicates.

When to Use It

Onboarding new predicates or rules in a domain and adding them with a clear domain, type, and signature
Preventing duplicates by checking exact name matches and semantic similarity before adding new items
Linking related predicates across domains to enable cross-domain queries
Exporting the registry to a file for backups or migrations
Searching the registry to discover existing predicates before reuse or modification

Quick Start

Step 1: /logic-register add --domain legal "canAccess(Person, Resource) ::: Person is authorized to access Resource"
Step 2: /logic-register search "access"
Step 3: /logic-register export --format json ./registry-backup.json

Best Practices

Always search before adding new predicates to avoid duplicates
Use descriptive, domain-conventional names and consistent arity
Include concrete examples in complex predicates to illustrate usage
Link related predicates to improve query resolution and discoverability
Regularly export backups of the registry for safety and recovery

Example Use Cases

Add a legal predicate canAccess(Person, Resource) with a clear description and signature
Search for predicates containing 'access' to find related access-control rules
List all registered predicates in the legal domain to review coverage
Link legal.canAccess with hr.hasEmploymentStatus to enable cross-domain checks
Export the registry to registry-backup.json for offline storage

Frequently Asked Questions

Add this skill to your agents