Public MCP server for Present Agent — find 5 explainable, personalized gift recommendations from any MCP client (Claude Code, Codex, and more). Hosted: no local catalog, Shopify, or model keys required.
Two tools with reasonable parameter coverage but significant gaps in descriptions and output documentation. present_find_gift has a detailed schema (10 params, all described) but lacks output schema documentation and error recovery guidance. present_beta_start is trivial (no params, minimal description). Descriptions are adequate (194 chars avg baseline met for find_gift) but output structure is never formally documented. No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Error handling delegates to raw API responses without recovery guidance.
Return the fastest beta onboarding instructions and live web fallback.
Find 5 personalized gifts through the hosted Present Agent API. Optional local Claude/Codex context search is opt-in.
Output schema not documented. present_find_gift returns JSON with fields like sessionId, picksUrl, contextSignals, but LLM cannot infer structure. Agents cannot plan downstream calls or extract data reliably.
Error handling returns raw API errors (status code + JSON payload) without recovery guidance. 'hosted_api_failed' tells agent nothing about retry eligibility, user-fixable issues, or next steps.
present_beta_start has minimal description (27 chars, below 34 char p10 baseline). 'Return the fastest beta onboarding instructions and live web fallback' lacks context for when/why to call it vs present_find_gift.
No tool annotations. Both tools are read-only (risk: READ_ONLY noted in metadata) but readOnlyHint not declared in schema. Agents cannot infer safety/idempotency without explicit hints.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 51 | 2026-07-28+ | v2 |
present_beta_start accepts no parameters but description implies it returns 'instructions and fallback'. No output schema means agent cannot extract URLs, steps, or structured guidance.