Tool fabric
Capability-aware tool selection, risk metadata, reliability and outcome learning.
SCHX / FEATURE LAYER
Source map
- python/abria_server/tool_fabric.py
- python/abria_server/intelligence.py
SCHX / FEATURE LAYER
Tool specs
ToolSpec records domains, capabilities, risk, parallel safety, workspace/network requirements, authority, reliability and cost.
SCHX / FEATURE LAYER
Examples
Default tools include personal memory search and reasoning-pattern retrieval; the catalog can also represent workspace-write, execute and external-network capabilities.
SCHX / FEATURE LAYER
Risk-aware routing
A tool plan can select the smallest reliable set of tools and expose why each tool matched. Risk is explicit rather than inferred after execution.
SCHX / FEATURE LAYER
Outcome feedback
The API includes tool-outcome reporting so tool reliability can be updated from observed results.