- Add AI Agent Governance tutorial: Policy-based sandboxing for single agents
- Teaches deterministic policy enforcement concepts
- Includes filesystem, network, and rate limit policies
- Full working Python code with audit logging
- Add Multi-Agent Trust Layer tutorial: Secure agent-to-agent communication
- Teaches trust scoring and behavioral monitoring
- Includes delegation chains with scope narrowing
- Full working Python code with audit trail
- Added app screenshots (01_home.png, 02_fullpage.png)
- Fixed HttpAgent import to use @ag-ui/client package
- Added .gitignore files for frontend and backend
- Updated package.json with correct @ag-ui/client version
A real-time agent vs agent negotiation showdown using Google ADK:
- Two AI agents (Buyer vs Seller) negotiate autonomously
- Dramatic used car scenario with secrets on both sides
- 4 buyer personalities (Desperate Dan, Analytical Alex, Cool-Hand Casey, Fair-Deal Fran)
- 4 seller personalities (Shark Steve, By-The-Book Beth, Motivated Mike, Drama Queen Diana)
- 3 negotiation scenarios (Used Car, Vintage Guitar, Apartment Sublet)
- Configurable max rounds and personality selection
- Live Streamlit UI showing offers, counteroffers, and outcomes
- Uses gemini-3-flash-preview model
Run with: streamlit run negotiation_app.py
Fixes#409
## Changes
### agent.py
- Replace singleton `google_search` with explicit `GoogleSearchTool(bypass_multi_tools_limit=True)`
to enable compatibility with nested agent setups (ADK limitation workaround)
- Update SearchAgent instruction to be clearer about how the built-in search works
(it's automatic, not a callable function)
### tools.py
- Add validation for common LLM hallucination patterns (e.g., `_v0` version numbers)
- Add fallback logic when artifact loading fails:
1. Try corrected filename if version 0 detected
2. Look up known artifact versions in session state
3. Fall back to last generated rendering
- Improve error messages to show available artifacts
## Root Cause Analysis
1. **google_search not found**: The LLM was trying to call 'google_search' as a
function, but GoogleSearchTool is a Gemini built-in capability that works
automatically. Using `bypass_multi_tools_limit=True` and clearer instructions
resolves the confusion.
2. **Artifact not found**: The LLM sometimes hallucinated incorrect filenames
(e.g., `_v0.png` instead of `_v1.png`). Added robust fallback logic to
recover gracefully.
- Changed response type to `RunOutput` in multimodal design and coding agent teams for improved type safety.
- Updated the `agno` package version in requirements.txt to ensure compatibility across all agents.
- Changed response type to `RunOutput` in the AI real estate and recruitment agent teams for improved type safety.
- Updated the `agno` package version in requirements.txt to ensure compatibility across all agents.
- Adjusted requirements for other dependencies to maintain consistency.
- Refactored legal agent team to utilize the new `Knowledge` class instead of `PDFKnowledgeBase`.
- Updated response type to `RunOutput` for better type safety across multiple functions.
- Changed agent model version from GPT-4.1 to GPT-5 for enhanced performance.
- Specified minimum version for the 'agno' package in requirements.txt for compatibility.
- Integrated ExaTools for competitor URL retrieval in the AI competitor intelligence agent.
- Updated response handling to utilize `RunOutput` for improved type safety.
- Specified minimum version for the 'agno' package in requirements.txt for compatibility across agents.
- Changed response type to `RunOutput` in various AI agents for improved type safety.
- Updated the `agno` package version in requirements.txt to ensure compatibility across all agents.
- Changed response type to `RunOutput` in both the 3D pygame and Tic Tac Toe agents for improved type safety.
- Updated the `agno` package version in requirements.txt to ensure compatibility.
- Introduced a new multi-agent system for analyzing landing page designs and providing expert UI/UX feedback.
- Implemented a Coordinator/Dispatcher pattern with specialized agents for visual analysis, design strategy, and visual implementation.
- Added tools for editing and generating improved landing pages, along with comprehensive reporting features.
- Introduced a multi-agent system for home renovation planning that analyzes photos and generates personalized renovation plans.
- Implemented a Coordinator/Dispatcher pattern with specialized agents for visual assessment, design planning, and project coordination.
- Added tools for generating and editing photorealistic renderings, along with budget-aware planning features.
- Updated the email generation workflow to yield progress updates during company processing.
- Refactored the database integration to use SqliteDb for better management.
- Enhanced the user interface with a more structured display of generated emails, company research, and contact information.
- Added troubleshooting tips in case of no email generation.
- Updated README with detailed features, installation instructions, and usage guide.
- Increased agno package version requirement to 2.0.4 for compatibility.
