Followr MCP server. Connects Claude and other AI agents to Followr's social media management platform. Manage posts, generate AI content, work with avatars and videos, and orchestrate social media workflows from natural language.
Four tools with complete input schemas and descriptions, but significant gaps in parameter detail and output documentation. Tool names use domain-specific verbs (weekly-brief, campaign-launch, video-series, crisis-response) but lack the standard verb_noun pattern (e.g., 'generate_weekly_brief'). Descriptions are adequate (100-200 chars) but lack actionable context for LLM selection. Parameters have types and basic descriptions, but many lack constraints (enums, ranges, format specs). No visible output schema documentation. Error handling is structured (tool-error.ts shows recovery guidance), but not integrated into tool definitions.
Launch a multi-network campaign end-to-end. Spin up a full campaign in one shot: hashtag/tag taxonomy, brand-aligned hero asset, teaser + launch + follow-up posts across selected networks, scheduled around the launch date.
Draft three crisis-response variants for review. Quickly produce three differently-toned crisis-response post drafts (apology, clarification, deflection) staged as drafts (not auto-publish) so a human can approve one.
Generate an avatar video series on a single topic. Produce N short avatar videos on a topic. One script per episode, one lipsync render per episode, all scheduled across N consecutive days.
Generate a week of scheduled posts from a brief. Take a free-form weekly brief and produce a full week of scheduled posts in the company. Anchors to the company's brand voice, picks suitable networks, drafts copy + images, and schedules across the week.
Tool names do not follow verb_noun pattern. 'followr.weekly-brief' should be 'followr.generate_weekly_brief' or 'followr.create_weekly_posts'. Current names are domain-specific but ambiguous to LLMs unfamiliar with Followr terminology.
No output schema documentation visible. Tools return structured results but LLMs cannot see what fields to expect. This breaks downstream tool chaining and forces agents to guess field names.
Parameters lack enums and format constraints. 'networks' accepts 'comma-separated' but no enum of valid values (instagram, facebook, tiktok, etc.). 'posts_per_day' is described as 'Integer; default 1' but no min/max bounds. LLMs will hallucinate invalid values.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 51 | 2025-06-18+ | v2 |
Descriptions lack dependency hints and selection context. 'Generate a week of scheduled posts' does not explain WHEN to use this vs campaign-launch or video-series. No guidance on prerequisites (e.g., 'company must have connected accounts').
tool-error.ts shows structured error handling with recovery guidance, but error responses are not documented in tool definitions. LLMs cannot see what errors to expect or how to recover.