Installable skillMIT

Ontological Renamer

Renames projects and content with dense, meaningful ontological titles that describe essence and function. Combines 3-4 words using separator conventions (- for compound/close words, -- for distant concepts). Provides translations to Latin and Greek. Use when naming projects, repositories, system...

BY luokai0UPDATED Oct 27, 2026
INTENDED FOR◈OpenClawInstall in a listed agent.

instructions.md

---
name: ontological-renamer
description: Renames projects and content with dense, meaningful ontological titles that describe essence and function. Combines 3-4 words using separator conventions (- for compound/close words, -- for distant concepts). Provides translations to Latin and Greek. Use when naming projects, repositories, systems, or concepts.
license: MIT
complexity: intermediate
time_to_learn: 30min
tags: [naming, ontology, taxonomy, branding, translation]
governance_phases: [shape]
organ_affinity: [all]
triggers: [user-asks-about-naming, user-asks-about-renaming, context:naming]
---

# Ontological Renamer

Create dense, meaningful names for projects, systems, and concepts using ontological principles and classical language translations.

## When to Use

- Naming new projects or repositories
- Renaming existing systems
- Creating brand identities
- Naming concepts, patterns, or frameworks
- Building taxonomies and vocabularies

## Naming Methodology

### Step 1: Identify Essence

What IS the thing? (noun/subject)

```
Questions:
- What category does it belong to?
- What is its fundamental nature?
- What metaphor best describes it?

Examples:
- A collection → repository, vault, archive
- A process → engine, forge, factory
- A connection → bridge, link, weave
```

### Step 2: Identify Function

What does it DO? (verb/action)

```
Questions:
- What transformation does it perform?
- What problem does it solve?
- What capability does it provide?

Examples:
- Creates → forge, builder, generator
- Connects → bridge, link, mesh
- Organizes → lattice, matrix, graph
```

### Step 3: Identify Domain

What CONTEXT does it operate in?

```
Questions:
- What field or industry?
- What system or ecosystem?
- What users or actors?

Examples:
- AI/ML → agent, model, neural
- Development → code, build, deploy
- Knowledge → wisdom, expertise, craft
```

### Step 4: Combine

Form a 3-4 word compound that reads like a dense sentence.

```
Pattern: [Domain]-[Function]--[Essence]-[Modifier]

Examples:
- agent-skill--knowledge-forge
- code-craft--pattern-vault
- data-weave--insight-engine
```

## Separator Conventions

### Single Hyphen (-)

Use for magnetically close concepts that form a compound:

```
skill-bundle      → "bundle of skills" (compound noun)
agent-powered     → "powered by agents" (adjective)
code-craft        → "craft of coding" (compound noun)
knowledge-base    → "base of knowledge" (compound noun)
```

### Double Hyphen (--)

Use for conceptual distance or category separation:

```
agent-skills--knowledge-engine
├── agent-skills   (what it contains)
└── knowledge-engine (what it is)

frontend-toolkit--design-system
├── frontend-toolkit (domain)
└── design-system (function)
```

### Decision Matrix

| Relationship | Separator | Example |
|--------------|-----------|---------|
| Adjective + Noun | - | `smart-agent` |
| Noun + Noun (compound) | - | `skill-bundle` |
| Category + Instance | -- | `dev-tools--linter` |
| Domain + Function | -- | `ml-ops--model-registry` |
| What + How | -- | `data--transform-engine` |

## Word Selection Rules

### Ignore These Words

When forming names, filter out:

```
Articles: a, an, the
Prepositions: of, for, to, in, on, by, with
Conjunctions: and, or, but
```

### Prefer Strong Words

```
Strong Nouns:
✓ forge, vault, engine, lattice, codex
✗ thing, stuff, helper, manager

Active Verbs:
✓ forge, weave, craft, build
✗ do, make, handle, process

Specific Terms:
✓ authentication, pipeline, schema
✗ security, flow, structure
```

### Domain-Specific Vocabulary

| Domain | Strong Words |
|--------|--------------|
| AI/Agents | agent, model, neural, inference, swarm |
| Development | code, build, deploy, pipeline, stack |
| Data | schema, query, transform, pipeline |
| Security | vault, shield, guard, sentinel |
| Knowledge | codex, archive, wisdom, expertise |

## Classical Translations

### Translation Approach

1. Identify root concepts in English
2. Find Latin/Greek equivalents
3. Combine following classical morphology
4. Verify pronunciation and semantic accuracy

### Common Roots

| English | Latin | Greek |
|---------|-------|-------|
| skill | ars, artis | technē |
| knowledge | scientia | epistēmē, gnōsis |
| agent | agens, agentis | praktōr |
| forge | fabrica | chalkeion |
| vault | camera | thalamos |
| codex | codex | biblos |
| mastery | magisterium | epistasia |
| bundle | fasciculus | desmos |
| craft | ars | technē |
| power | potestas | dynamis |

### Translation Examples

```
skill-codex--agent-mastery
├── Latin: artium-codex--agentis-magisterium
└── Greek: technōn-biblos--praktoros-epistasia

agent-skill--knowledge-forge
├── Latin: ars-agentis--scientia-fabrica
└── Greek: technē-praktor--epistēmē-chalkeion
```

## Name Generation Process

### Example: AI Skills Repository

**Context:** A repository of skills that extend AI agent capabilities.

**Step 1: Essence**
- It's a collection (repository, vault, archive)
- It contains skills (capabilities, expertise)
- It's curated (codex, library)

**Step 2: Function**
- Extends capabilities (augment, enhance)
- Provides knowledge (expertise, mastery)
- Bundles together (bundle, collection)

**Step 3: Domain**
- AI agents (agent, autonomous)
- Skills/capabilities (skill, craft, art)
- Knowledge systems (knowledge, wisdom)

**Step 4: Generate Candidates**

```
1. agent-skill--knowledge-forge
   "A forge of knowledge for agent skills"

2. skill-codex--agent-mastery
   "A codex of skills for agent mastery"

3. capability-weave--agent-loom
   "A loom that weaves agent capabilities"

4. praxis-bundle--agent-expertise
   "A bundle of practical expertise for agents"

5. craft-matrix--agent-powers
   "A matrix of crafts that power agents"
```

**Step 5: Evaluate**

| Name | Clarity | Memorability | Domain Fit |
|------|---------|--------------|------------|
| skill-codex--agent-mastery | High | High | High |
| agent-skill--knowledge-forge | High | Medium | High |
| capability-weave--agent-loom | Medium | Medium | Medium |

**Winner:** `skill-codex--agent-mastery`

### Example: Authentication Library

**Context:** A library for handling user authentication.

**Candidates:**

```
1. auth-guard--access-sentinel
   "A sentinel guarding access through authentication"

2. identity-forge--credential-vault
   "A vault of credentials from an identity forge"

3. login-shield--session-keeper
   "A shield and keeper of login sessions"
```

## Output Format

When generating names, provide:

```yaml
naming_proposal:
  context: "Brief description of what's being named"

  candidates:
    - name: "english-compound--name"
      meaning: "What this name conveys"
      latin: "latin-translation"
      greek: "greek-translation"
      pros: ["list", "of", "strengths"]
      cons: ["list", "of", "weaknesses"]

  recommendation:
    name: "recommended-name"
    rationale: "Why this is the best choice"
```

## Anti-Patterns

### Avoid These

```
❌ Generic Names
   my-project, cool-tool, awesome-thing

❌ Overloaded Terms
   manager, handler, service, util

❌ Acronyms Without Meaning
   XYZPT, QWERT

❌ Too Long
   super-advanced-machine-learning-based-intelligent-agent-system

❌ Unclear Separators
   agent_skill-knowledge--forge (mixing _ and -)
```

### Prefer These

```
✓ Evocative Metaphors
   forge, vault, weave, lattice

✓ Domain-Specific Terms
   codex, praxis, schema

✓ Clear Compound Structure
   skill-bundle, knowledge-forge

✓ Balanced Length (2-5 words)
   agent-skill--knowledge-forge
```

## References

- `references/naming-patterns.md` — Compound naming patterns
- `references/separator-conventions.md` — Detailed separator rules
- `references/translation-guide.md` — Latin/Greek translation guidance
- `references/workflow-integration.md` — Ecosystem integration
- `assets/word-roots.md` — Common Latin/Greek roots

## Related Skills

- **documentation-generator**: For documenting named concepts
- **brand-guidelines**: For visual identity to match names
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.