Overview
Gives the Agent an explicit scaffold for thinking — decompose, list assumptions, backtrack and verify. Good for multi-step decisions where the reasoning must be auditable.
Features
- Explicit steps, easy to audit
- Supports backtracking and branching
- Composable with other tools
- Exports a readable reasoning trace
- Lightweight, no external dependencies
Exposed tools
Once connected, the Agent can call these tools directly in conversation
sequentialthinking
