Loading ToolBench
A tamper-evident record of what an agent actually did — hash-chained, signed, anchored, and honest about what that does not prove.
This is a test/demo MCP server, not a production tool. All three tools have minimal descriptions (6-27 chars, well below the 34-char p10 baseline) and lack parameter documentation. The 'echo' tool has a basic schema with one string param but no description for that param. 'explode' and 'mutate' have empty input schemas (type:object with no properties). No error handling guidance, no output schemas documented, no tool composition or chaining support. This server exists to demonstrate acta's ledger recording, not to showcase tool quality.
echo
always fails
announces that the definitions changed
All tool descriptions are under 20 characters (echo: 4, explode: 12, mutate: 33). LLMs cannot determine when or why to select these tools.
Two tools (explode, mutate) have empty input schemas: {type:object} with no properties defined. LLMs cannot understand what parameters these tools accept.
The 'echo' tool's 'text' parameter has no description. LLMs cannot infer parameter meaning from names alone.
No output schemas documented for any tool. LLMs need to know what fields to expect in responses to plan downstream calls and extract data correctly.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 32 | 2026-07-28+ | v2 |
No error handling guidance. The 'explode' tool always fails but provides no recovery hint (e.g., 'This tool is designed to fail; use echo instead'). Error responses must tell the LLM what to do next.