According to AlphaMoat, as of 2026-09, ontology-clawra has 1,784 downloads and 1 stars — ranked #4,586 of 5,078 AI skills overall, and #902 of 969 in AI Agent.
| Platform | Installs | Downloads | Stars | Comments |
|---|---|---|---|---|
| ClawHub ↗ | 30 | 1.5K | 1 | - |
| SkillHub ↗ | 215 | 1.8K | 1 | — |
Platforms count differently and update at different times.
The skill describes structured knowledge storage, confidence calculation, rule learning, and reading workspace memory. File access is limited to the documented paths; it explicitly excludes cross-skill scheduling and automatic network requests. This does not mean local files were read or changed. Ontology files may store persistent workspace context; inspect proposed writes and keep sensitive information out unless its storage is authorized.
Profile published 2026-10-27
Generated from the skill’s documentation as examples; not results of actual runs.
The counting method changed in 2026-09: skills with the same name are now counted separately by author, and the full catalog is covered. Dashed lines show months before the change; they are not compared with later months, and no growth is computed across the change.
| Month | Downloads | MoM | Stars |
|---|---|---|---|
| 2026-09 | 1.8K | — | 1 |
| 2026-08 | 1.4K | +8.3% | 1 |
| 2026-07 | 1.3K | +13.9% | 1 |
| 2026-06 | 1.1K | +9.1% | 1 |
| 2026-05 | 1.0K | — | 1 |
| 2026-03 | 473 | — | - |
The author’s original file, shown as plain text without edits or review by AlphaMoat. Links are not clickable; check commands before running them. Fetched 2026-10-23.
--- name: ontology-clawra description: | ontology-clawra v5.0 - Local Ontology Reasoning Engine A local structured reasoning engine with confidence management and automatic learning. Reads user workspace memory files for context. No cross-skill scheduling or network communication. Core Capabilities: - Ontology knowledge graph storage and reasoning
- Confidence calculation and meta-cognition
- Local memory search and rule learning
File Access: Read/write only ~/.openclaw/skills/ontology-clawra/memory/ for ontology files;
Read only ~/.openclaw/workspace/memory/*.md for user context.
No cross-skill scheduling, no automatic network requests.
metadata:
{
"openclaw": {
"emoji": "🧠",
"version": "5.1.0",
"last_updated": "2026-03-25",
"changelog": [
"v5.0.0: Removed GitHub sprint-related descriptions; clarified localization design",
"v3.8.0: Added Skill ecosystem documentation",
"v3.7.0: Enabled automatic learning"
]
}
}
---
# ontology-clawra v5.0 - Local Ontology Reasoning Engine
## 1. Core定位 (Accurate Description)
**This is a local reasoning engine, not a cross-skill coordinator.**
```
Actual Capabilities:
✅ Local ontology knowledge graph storage and reasoning
✅ Confidence calculation and meta-cognition
✅ Local memory file search
✅ Automatic learning (write to local memory directory)
Actual Limitations:
❌ No skill scheduling (no Event Router)
❌ No cross-skill communication
❌ No automatic network requests
❌ No agent-to-agent RPC
```
## 2. File Access Permissions
| Operation | Path | Description |
|-----------|------|-------------|
| Read | `~/.openclaw/skills/ontology-clawra/memory/` | Ontology files |
| Read | `~/.openclaw/workspace/memory/*.md` | User context |
| Write | `~/.openclaw/skills/ontology-clawra/memory/` | Learning results |
| Write | GitHub (via git push) | Requires user confirmation |
| ❌ | Other directories | Never accessed |
**Network Behavior**:
- No automatic network requests
- network_fetch.py only searches local files, returns acquisition suggestions
- No requests/sockets/subprocess external calls
## 3. Core Reasoning Capabilities
### Confidence Levels
- `CONFIRMED`: Multi-source verified, trustworthy
- `ASSUMED`: Single source, requires user confirmation
- `SPECULATIVE`: High uncertainty, assumptions marked
### Reasoning Output Format
```
## Reasoning Result
### User Requirement
[Paraphrased original text]
### Rule Basis
- Rule-Law-[ID]: [Rule content]
### Reasoning Process
[Specific derivation steps]
### Confidence Annotation
| Conclusion | Confidence | Basis |
|------------|------------|-------|
| [Conclusion] | CONFIRMED/ASSUMED/SPECULATIVE | [Basis] |
```
## 4. Ontology File Format
**Directory**: `~/.openclaw/skills/ontology-clawra/memory/`
| File | Purpose |
|------|---------|
| `graph.jsonl` | Entities (Concept/Entity) |
| `rules.yaml` | Rules (Rule/Law) |
| `laws.yaml` | Patterns (inductive patterns) |
| `confidence_tracker.jsonl` | Confidence tracking |
| `reasoning.jsonl` | Reasoning log |
## 5. Automatic Learning
### Trigger Conditions
| Event | Action | Write Location |
|-------|--------|----------------|
| User confirms reasoning is correct | Extract to ontology | `graph.jsonl` |
| Confidence can be upgraded | Update confidence | `confidence_tracker.jsonl` |
| Reasoning fails | Suggest ontology addition | Prompt only |
### Write Rules
- Inform user before all write operations
- Write only to `~/.openclaw/skills/ontology-clawra/memory/`
- Do not upload any data to external services
## 6. Security Statement
1. **No external network calls**: Code contains no requests/urllib/sockets external calls
2. **No credential access**: Does not read environment variables or API keys
3. **Privacy isolation**: Workspace memory files used only for context, not synced
4. **Localization**: All learning results stored only in local memory directory
## 7. Version History
| Version | Date | Description |
|---------|------|-------------|
| v5.0 | 2026-03-25 | Removed GitHub sprint-related descriptions |
| v3.9 | 2026-03-24 | Removed exaggerated descriptions, clarified localization |
| v3.8 | 2026-03-20 | Added Skill ecosystem documentation |
| v3.7 | 2026-03-16 | Enabled automatic learning |
| v3.6 | 2026-03-16 | Initial version |
## 8. Supported Domain Ontologies
Supply chain procurement, healthcare, finance banking, cybersecurity,
automotive manufacturing, human resources, photography, furniture,
elderly care, childcare, tea culture, moving services, dating,
aerospace, and 50+ domains.
Ontology files located in: corresponding yaml files under `~/.openclaw/skills/ontology-clawra/memory/`Records learnings, errors and corrections as Markdown files for later reuse and promotion into workspace memory.
Finds skills in local and remote catalogs from a plain-language request and recommends what to install.
A browser-automation command line skill for AI agents.
Turns corrections and self-review into tiered memory and promotes or archives it by use.
Reviews a third-party skill's source, permissions, and danger patterns against a checklist before it is installed.
Stores agent memory and cross-skill state as a typed, constraint-checked knowledge graph.
A workspace setup guide for planning proactive follow-up, working memory and review routines for an AI agent.
A guide to creating and updating effective Claude skills.
Data month: 2026-09 · Downloads, stars and installs are aggregated monthly from public skill registries (ClawHub, SkillHub). See methodology.