MCP server for cross-posting content to Telegram, Discord, Mastodon, Bluesky, and Reddit
Three tools with clear verb-noun naming (post_to_*) and reasonable descriptions (120-180 chars each). All parameters have type definitions and descriptions. However, schemas lack critical details: no output schema documentation, no error handling guidance, no parameter constraints (e.g., max length validation), and no idempotency markers. The post_to_all tool combines multiple platform operations but lacks per-platform error recovery guidance. Descriptions mention platform limits but don't guide LLM on handling failures. Missing: enum constraints for status values, structured error recovery paths, and output field documentation.
Post the same message to all configured platforms concurrently. Runs all platform posts in parallel. Individual platform failures do not abort other platforms; each result is reported separately. Args: text: Message text. Note that platforms have different length limits (Discord 2000, Telegram 4096). Text exceeding a platform's limit will fail for that platform only. Returns: Dict with overall status and per-platform results: {"status": "ok" | "partial" | "error", "results": [...]}
Post a message to the configured Discord channel via webhook. Args: text: Message content (max 2000 chars). Supports Discord markdown: **bold**, *italic*, `code`, ```block```. Returns: Dict with platform, status, message_id, channel_id.
Post a message to the configured Telegram channel. Args: text: Message text (max 4096 chars). Supports HTML formatting: <b>bold</b>, <i>italic</i>, <code>mono</code>, <a href="">link</a>. Returns: Dict with platform, status, message_id, chat_id.
No output schema documentation. Tools return dict[str, Any] with undocumented fields (platform, status, message_id, chat_id, error, error_type). LLMs cannot plan downstream operations or extract required fields without guessing.
No error handling guidance. Descriptions mention platform limits (Discord 2000, Telegram 4096) but don't tell LLM what to do on failure. No recovery hints like 'If text exceeds limit, try post_to_telegram instead' or 'Retry with shorter text'.
post_to_all lacks per-platform failure semantics. Description says 'Individual platform failures do not abort other platforms' but doesn't explain what LLM should do with partial results. Should it retry failed platforms? Notify user? Continue?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | 2026-07-28+ | v2 |
No idempotency markers. These are WRITE tools (risk=WRITE) that post to external platforms. No indication whether repeated calls with same text create duplicate posts or are idempotent. Agents retry on ambiguous failures, risk of duplicate posts.
Parameter 'text' lacks validation constraints in schema. Descriptions mention max lengths (2000, 4096) but JSON Schema has no maxLength constraint. LLMs cannot validate before calling; will pass oversized text and fail.