Multi-agent mathematical discovery engine with BDI architecture. Handles project-specific file organization, archiving rules, agent hierarchy, and testing workflows for the SYMBO_AGENTIC_REASONERS codebase.
A development assistant skill for the SYMBO_AGENTIC_REASONERS project (v0.6.0), a multi-agent mathematical discovery engine built with BDI (Belief-Desire-Intention) architecture.
SYMBO_AGENTIC_REASONERS is a sophisticated multi-agent system designed for mathematical problem-solving and discovery. The project uses a hierarchical agent architecture with domain-specific supervisors and task specialists.
Follow these strict file location conventions:
- Current: `docs/assessments/critical/`
- Archived: `docs/assessments/archive/`
When archiving assessments or audits, follow this three-step process:
1. **Copy to archive** with naming pattern: `{NAME}_v{VERSION}_{YYYY-MM-DD}.md`
2. **DELETE the original** after confirming the archive exists
3. **Never leave duplicates** in both locations
Archive locations:
**Always use**:
```python
from symbo_agentic_reasoners.{module} import ...
```
**Never use** legacy phase folder imports.
The system uses a three-tier agent architecture:
- Algebra
- Calculus
- Linear Algebra (LinAlg)
- Statistics (Stats)
Run tests using either method:
```bash
python -m pytest tests/
python scripts/quick_test.py
```
The configured test command runs pytest with fail-fast mode (`-x` flag).
Linting is configured with Ruff:
```bash
ruff check --fix
```
Always reference the project context file when working:
Leave a review
No reviews yet. Be the first to review this skill!
# Download SKILL.md from killerskills.ai/api/skills/symbo-agentic-reasoners-development-assistant/raw