Loading ToolBench
An architecture modeling and failure simulation platform for reviewing system designs and testing resilience. Provides tools for building system models, running failure scenarios, and collaborating on architecture decisions.
Aether has 18 tools with minimal schema documentation. Most tools lack visible input parameter schemas in the source code, only 2 tools (run_failure_scenario, propose_architecture_change) show explicit parameter definitions. Descriptions are present but extremely terse (5-30 chars), far below the 50-200 char baseline for LLM-optimized descriptions. No output schemas documented. Tool names follow verb_noun convention (get_, create_, read_, etc.), which is good, but parameter descriptions are absent for 16 of 18 tools. Error handling, recovery guidance, and security patterns are not evident in the provided code. The server appears to be a React/Hono frontend with HTTP transport, but tool registration and schema enforcement are not visible in the source excerpts provided.
Add a component
Leave a note on the record
Compare the futures
Connect two components
Open a repair future
Read the architecture
Read the decision record
Inspect what a failure reaches
16 of 18 tools have no visible input parameter schemas. Only run_failure_scenario and propose_architecture_change show explicit parameter definitions. This violates the critical requirement that every tool must declare its input schema with types and descriptions.
All 18 tool descriptions are under 30 characters (baseline 50-200 chars for LLM optimization). Descriptions like 'Read the decision record' and 'Join the incident room' lack context for when/why to use the tool, prerequisites, and what the tool returns. This forces LLMs to guess at tool selection.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 31 | 2026-07-28+ | v2 |
Join the incident room
Measure a dependency's real demand
Build a system from your description
Propose a change to a future
Read a component's traffic
Read a live status source
Read a repository's architecture
Recommend one, and say why
Simulate a failure
Trace a dependency
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields without knowing what each tool returns. This violates the requirement that tools must document their return types.
No error handling or recovery guidance visible. Tools do not indicate what errors are retryable, user-fixable, or fatal. No actionable error messages documented. This prevents agents from self-correcting on failures.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) visible in source. Risk field is provided in the evaluation metadata but not integrated into tool definitions. This prevents agents from understanding side effects and retry safety.