Delegate a coding task from any MCP client (Cursor, VS Code, Windsurf, Zed, Claude Code, CI) to an isolated microsandbox microVM and run Claude Code inside it.
Critical gaps across all dimensions. Tool definitions lack input schemas entirely, cannot verify parameter types, constraints, or descriptions. No visible schema registration in src/handlers.ts. Tool descriptions are present but minimal (10-50 chars), providing insufficient context for LLM selection. No evidence of output schema documentation. Error handling and recovery guidance absent. Composition is reasonable (5 focused tools), but without schemas and detailed descriptions, agents cannot reliably invoke these tools.
Delegate a coding task to an isolated microsandbox microVM
Get the status of the microsandbox pool
Resume a delegated task with additional input
Get the status of a delegated task
Teardown a delegated task and clean up resources
No input schemas visible for any tool. Cannot verify parameter types, constraints, enums, or descriptions.
Tool descriptions are too brief (10-50 chars). Rubric baseline: 194 chars average, p10=34, p90=392. Current descriptions lack WHAT the tool does, WHEN to use it, and prerequisites.
No output schema documentation. LLMs cannot plan downstream calls or extract required fields. Rubric: 100% of A+ tools have documented return types.
No error handling guidance. Errors should tell LLMs what to do next (retryable? user-fixable? fatal?). Currently absent.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 33 | <=2025-11-25 | v2 |
Tool definitions inferred from handler names only; no explicit schema registration visible in source.