feat: implement local SemanticSearchAgent for .ics parsing and CalDAV integration

[anirudhagarwal-dev]

Summary

This PR implements the Semantic Search Agent, adding background local file indexing and natural language search capabilities as requested in issue #289.

Changes made

• Created a new SemanticSearchAgent in daemon/pilot/agents/semantic_search_agent.py
• Added SemanticSearchConfig in config.py for configurable folders and index storage
• Added AgentRole.SEMANTIC_SEARCH in base_agent.py
• Updated READ_ONLY_ACTIONS in actions.py to include workspace indexing
• Uses the existing WorkspaceIndex which already handles FAISS and sentence-transformers!

Type of change

• New feature / enhancement
• Bug fix
• Documentation update
• Tests / CI
• Performance improvement
• Refactor (no functional change)

Closes

Closes #289

[VyomKulshrestha]

Hi @anirudhagarwal-dev , thanks for working on the Semantic Search Agent! The integration with the @auto_register decorator and the new WORKSPACE_INDEX / WORKSPACE_SEARCH action types looks very clean.

However, it looks like you forgot to commit the actual workspace_index.py file! The SemanticSearchAgent explicitly imports from pilot.memory.workspace_index import WorkspaceIndex and calls its methods, but that file isn't included in the pull request diff.

If we merge this right now, the daemon will instantly crash on startup with a ModuleNotFoundError: No module named 'pilot.memory.workspace_index'. Could you please commit and push that file, along with any necessary vector database dependencies (e.g. Chroma, FAISS) to pyproject.toml? Thanks!