Zero's tool definitions are severely underdeveloped. Only 2 of 7 tools have visible input schemas (apply_patch, read_file); the remaining 5 lack documented schemas entirely. Descriptions are minimal (10-50 chars) and lack actionable context. No parameter descriptions are present. Tool definitions appear inferred from test code rather than formally registered. No output schemas documented. Error handling and recovery guidance absent. This server does not meet baseline quality standards for production agent tooling.
Applies patches to files in the workspace
Requests user input during execution
Lists files matching a glob pattern
Searches for patterns in files
Reads file contents from the workspace
Requests permission to execute sensitive operations
Searches for available tools by capability
5 of 7 tools have NO visible input schemas. Grep, glob, ask_user, request_permissions, tool_search lack any documented parameter structure.
All tool descriptions are under 50 characters and lack actionable context. Current descriptions are generic placeholders.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 33 | 2026-07-28+ | v2 |
No output schemas documented for any tool. LLMs need to know what fields to expect to plan downstream calls and extract data.
Tool definitions appear inferred from test code (freeform_tool_test.go) rather than formally registered.