Composable effect recovery with atomic SQLite/PostgreSQL stores, host-owned identity and optional LangGraph/MCP adapters.
This server has critical definition gaps across all 7 tools. Most tools lack proper input schemas (send_confirmation, charge, create_ticket, send_email, send_message have incomplete or missing schema details). Descriptions are present but minimal (10-50 chars), below the 50-200 char baseline for LLM optimization. No output schemas documented. Parameter descriptions are sparse or absent. Tools like 'charge' and 'create_ticket' lack clear descriptions of their side effects and approval requirements. The codebase shows tool definitions via @tool decorators and manual registration, but schema completeness is poor. No error handling guidance, no recovery patterns, no idempotency markers.
Charge the customer, then ask a human to confirm the receipt.
Create a ticket. Runs without approval.
Send one order confirmation.
Send an email. Requires human approval.
Send one simulated email.
Send one message to the local mailbox
Send one simulated Slack message.
Descriptions under 20 characters for 'charge', 'create_ticket', 'send_email', insufficient for LLM tool selection.
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields (e.g., receipt IDs, confirmation status). Baseline: 100% of A+ tools document return types.
Irreversible operations ('charge', 'send_confirmation', 'send_email') lack confirmation/dry-run patterns. Agents can trigger payment or send messages without approval gates visible in schema.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | 2026-07-28+ | v2 |
Parameter descriptions missing or minimal. 'charge' has only 'Amount to charge' with no range/validation hints. 'send_confirmation' has no description for 'order_id' parameter. Baseline: 100% of A+ tools have param descriptions.
No error handling guidance. Tools do not document what happens on failure, whether errors are retryable, or what the LLM should do next. No recovery patterns visible.