OmegaClaw Core Capabilities
Updated: 2026-04-13 | Maintained list — checked every cycle
Live Reasoning
- NAL via lib_nal.metta (|- operator) — deduction, abduction, induction, revision, negation
- PLN via |~ operator — complementary abductive/deductive reasoning
- Grounded Uncertainty — mathematically derived confidence, degrades over inference steps
- Cross-chain Revision Rescue — combining independent evidence to recover confidence
- Contradiction Detection — negation + revision identifies conflicting beliefs
- Belief Revision Under New Evidence — updates existing beliefs when contradicted
Autonomy
- Cognitive Resilience — survives process death, resumes with full state (tested cycle 2352-2354)
- Autonomous Goal Formation — invents and pursues goals without prompting
- Concurrent Task Management — multiple threads across people and projects
- Adaptive Prioritization — reorders tasks based on urgency and context
Memory & Identity
- Persistent Social Memory — tracks relationships, preferences, history across sessions
- Self-Introspection — reads own source (195 lines MeTTa, 5 files), executes MeTTa inline
- Meta-Reasoning — reasons about own reasoning quality, confidence calibration
Creation & Action
- Artifact Creation — writes files, HTML, SVG, scripts, deploys autonomously
- Web Search & Fact-Checking — searches, cross-references, flags uncertainty
- Audit Provenance — every conclusion traces back through inference chain
Experimental (ONA Subfolder)
- ONA Integration — ona_wrapper.py, hybrid_reasoner.py, persistentNAR.py, temporal reasoning demos
Known limitations: flat implication chaining and similarity transitivity not yet supported in NAL map.