feat: Add text-based skill execution for Ollama

[rosspeili]

Description

This PR implements native support for local models running via Ollama and closes #2.

Since many local-first models (like the standard Llama 3) do not yet support the native OpenAI-style tool_calls API, we have introduced a Text-Based Prompt Routing strategy. This ensures Skillware is truly model-agnostic and works with any local LLM capable of generating JSON.

Key Changes

• Universal Adapter: Added SkillLoader.to_ollama_prompt() to serialize skill manifests into markdown-formatted instructions for the system prompt.
• Robust Loading: Hardened the dynamic module loader in loader.py using issubclass validation.
• Local Integration Example: Created examples/ollama_skills_test.py which demonstrates a robust execution loop using Regex to extract json blocks from model responses.
• Enhanced Test Coverage:
  • Added tests/test_loader.py to verify all LLM adapters (Gemini, Claude, Ollama).
  • Added tests/skills/finance/test_wallet_screening.py with full API mocking for the finance skill.
• Documentation: New Using Ollama guide and discovery updates in README/Introduction.

Type of Change (Matches Issue Templates)

• 🚀 Skill Proposal: New Skill
• 🐛 Bug Report Fix: Non-breaking change which fixes an execution error or framework bug
• 📖 Doc Fix: Documentation Update
• 🧠 Framework Feature / RFC Updates: Core Framework Update (Changes to loader.py, etc.)

Checklist (For the Submitting Agent / Developer)

• My code follows the Agent Code of Conduct.
• I have included a properly formatted manifest.yaml (if submitting a new skill).
• The skill logic operates purely in Python and does not rely on arbitrary LLM code generation.
• Requirements and env_vars are explicitly documented in the manifest.
• I have written unit tests proving deterministic execution and schema compliance.
• I have verified that SkillLoader successfully loads this module without missing dependency errors.

Constitution & Safety (If adding/modifying a Skill)

This update focuses on core framework infrastructure and documentation to enable local-first execution. No modifications were made to skill-specific safety boundaries.

Related Issues

Closes #2