Remote MCP server for truval.dev — email verify and management API tools
Server has well-structured tool definitions with clear naming (verb_noun pattern), comprehensive descriptions (avg 180 chars), and complete JSON Schema with required/optional fields. All 8 tools are explicitly registered with inputSchema. However, output schemas are entirely undocumented, LLMs cannot predict response structure for downstream chaining. Parameter descriptions are present but lack constraint details (ranges, formats, enums). Error handling guidance is absent. Security-sensitive tools (create_api_key, revoke_api_key) lack permission/scope declarations. No tool annotations (readOnlyHint, destructiveHint) despite clear risk levels.
Create a new standard API key (sk_live_...) using your provisioning key (sk_mgmt_...). Returns the raw key once — store it securely.
Create a Stripe billing portal session URL for this account.
Get account tier, limits, billing status, and usage_this_month.
Get daily usage breakdown for the last 30 days.
Get month-to-date usage summary (calls used/remaining, estimated cost, tier). On free tier, calls_remaining is monthly verification quota headroom (aligned with verify API), not billable-only counts. On paid tiers, counts are billable (included allowance vs Stripe meter).
List all active standard API keys for this account (never returns raw key values).
Output schemas completely undocumented. LLMs cannot predict response structure (fields, types, nesting) for any tool. Breaks tool chaining, agent cannot extract IDs or references for downstream calls.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels. create_api_key and revoke_api_key are WRITE/DESTRUCTIVE but unmarked. LLMs cannot distinguish safe reads from dangerous mutations.
Parameter descriptions lack constraint details. 'email' has no format/validation hint. 'return_url' says 'must be a valid URL' but no regex/pattern. 'webhook_secret' requires 'min 8 characters' but not enforced in schema. LLMs cannot validate inputs before calling.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 68 | 2026-07-28+ | v2 |
Revoke a standard API key by id. Takes effect immediately.
Verify if an email address is real and deliverable. Returns valid, status, ordinal confidence (0-1, not a probability), and signals (disposable, role, catch_all, smtp_blocked, typo suggestion). Gmail/Outlook/Yahoo: smtp_blocked true, confidence ~0.75 — expected. Catch-all: confidence ~0.65; do not treat valid as mailbox proof. Unknown with MX but not smtp_blocked: SMTP inconclusive (e.g. org/M365), confidence ~0.50 — ask user to confirm, do not auto-reject. Limits: per-minute rate limits by tier; free tier also has 500 verification units per UTC calendar month (count_usage_for_customer; throttle/monthly-cap 429 rows logged but not counted). Paid overage uses billable units (post-syntax).
No error handling guidance. Descriptions do not explain failure modes, recovery steps, or retryability. E.g., verify_email mentions rate limits and monthly quotas but no guidance on 429/quota-exceeded responses.
Security-sensitive tools lack permission/scope declarations. create_api_key and revoke_api_key should declare required scopes (e.g., 'keys:write', 'keys:revoke'). No audit trail guidance for destructive operations.