andy-stack_vaultkeeper-ai/AIPrompts
Andrew Beal 32050767ca feat: enhance conversation naming with prompt injection protection and improved validation
- Wrap user messages in <message_to_title> tags to prevent prompt injection
- Expand NamePrompt with security guidelines and edge case handling
- Add 10-word limit truncation for generated titles
- Move conversation save before naming to ensure persistence
- Add comprehensive task complexity assessment framework to SystemPrompt
- Change stack limit error to warning log for better resilience
2026-01-01 13:06:06 +00:00
..
IPrompt.ts refactor: restructure AI prompt and agent architecture for multi-agent planning support 2025-12-30 19:07:00 +00:00
NamePrompt.ts feat: enhance conversation naming with prompt injection protection and improved validation 2026-01-01 13:06:06 +00:00
PlanningAgentSystemPrompt.ts refactor: improve function call/response parsing and filtering 2025-12-31 22:56:22 +00:00
SystemPrompt.ts feat: enhance conversation naming with prompt injection protection and improved validation 2026-01-01 13:06:06 +00:00