Loading ToolBench
Servidor Model Context Protocol para Penca Antel Ovación (no oficial) — jugá la penca de Antel Ovación desde tu asistente: pronosticá partidos, sumate a grupos, mirá el ranking, leé y publicá en el muro, y consultá los pronósticos con IA de Ovi.
Penca Ovación has 17 tools with consistent naming (penca_* prefix) and basic descriptions, but critical gaps undermine quality. Most tool descriptions are 20-80 chars, below the 194-char baseline for A+ tools. Parameter descriptions are sparse or missing entirely: penca_update_profile has 3 params with null descriptions; penca_predict lacks descriptions for homeScore/awayScore; penca_wall_post omits content/groupId descriptions. No output schemas are documented, LLMs cannot infer what fields to expect from responses. Error handling is minimal (guard wrapper catches exceptions but provides no recovery guidance). The server lacks tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite having clear WRITE vs READ_ONLY risk classifications. Pagination is present in some tools but inconsistently applied. Overall, the server is functional but falls short of production-grade tool design.
List news articles.
Ovi's daily AI digest.
Groups you belong to.
Public/featured groups you can join.
Sign out: revoke the Penca session server-side and clear the stored tokens for this connection.
Aggregate prediction statistics for a match.
List matches for a tournament (view: upcoming|finished).
No output schemas documented. LLMs cannot infer response structure, forcing them to guess what fields are available for downstream tool calls or data extraction.
Parameter descriptions are missing or null in 8+ tools. penca_update_profile has 3 params with null descriptions; penca_predict lacks descriptions for homeScore/awayScore; penca_digest has null description for 'kind' param.
Tool descriptions are too brief (20-80 chars vs 194-char baseline). Descriptions like 'List available tournaments' and 'List active polls' lack context for when/why to call them or what structure they return.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
Ovi's AI score prediction and reasoning for a match.
List active polls.
Submit (or overwrite) your score prediction for a match.
A user's predictions and stats (defaults to yourself).
Leaderboard for a group.
List available tournaments.
Update your profile (nickname, full name, country). Returns the updated account.
Publish a post to a group wall.
Read social wall posts, optionally scoped to a group.
Show the authenticated Penca account.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear WRITE vs READ_ONLY risk classifications. LLMs cannot distinguish safe reads from destructive writes without explicit hints.
Error handling provides no recovery guidance. The guard() wrapper catches errors and returns them as text, but does not categorize them as retryable/user-fixable/fatal or suggest next steps.