Loading ToolBench
MCP server for Follow Up Boss CRM integration with role-based access control (closer vs. manager). Provides tools for call logs, notes, lead search, team activity, and pipeline management.
Strong foundation with clear naming conventions (verb_noun pattern), comprehensive parameter schemas with enums, and good descriptions. All 12 tools have explicit definitions with input schemas visible in src/tools-manifest.ts. However, output schemas are not documented in the source code, and descriptions lack depth on error conditions and recovery paths. Tool composition is clean (single responsibility per tool), and access control is properly scoped (personal vs manager tools). Missing: documented return types, error handling guidance, and parameter relationship documentation.
Full activity for ONE closer for a period: their calls (with AI-summary notes), notes, and texts on touched leads. Accepts a name or a FUB user id. Manager only.
Your End-of-Day bundle for a period: YOUR calls (with AI summaries), the notes YOU wrote, and YOUR texts on the leads you contacted. All scoped to you. Paste your report template and Claude summarizes the key comms/updates from this data.
Search ALL Follow Up Boss leads by name (any assignee). Returns id, name, stage, assigned closer. Manager only.
Search YOUR assigned Follow Up Boss leads by name. Only leads assigned to you are returned.
Get the full timeline (calls, notes, texts) for one of YOUR leads. Refused if the lead is not assigned to you.
Output schemas not documented. No visible return type definitions for any of the 12 tools. LLMs cannot plan downstream operations or extract required fields without knowing response structure.
Error handling guidance missing. Descriptions do not explain what errors can occur, when to retry, or how to recover. E.g., 'Refused if the lead is not assigned to you' (get_lead_activity) lacks actionable recovery steps.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | B | 79 | 2026-07-28+ | v2 |
Full timeline (calls, notes, texts) for ANY lead by id. Manager only.
List YOUR Follow Up Boss calls for a period, each with the AI-summary note (score/topics/sentiment). Only your own calls are ever returned.
List the notes YOU authored on leads for a period. Only your own notes are returned.
List all Follow Up Boss team members (id, name, email, role). Manager only.
Exact numeric summary of YOUR calls for a period: total, answered, not-answered (approx), talk time, distinct leads.
TEAM leaderboard for a period: per-closer calls (total/answered/no-answer/talk time/distinct leads) and notes authored, plus a no-activity list. Manager only. (Texts are not counted team-wide — FUB has no account-wide by-agent text feed.)
Live pipeline: lead counts by stage x closer (active seller-deal stages only; excludes archived/closed/list/nurture buckets). Manager only. May take a few seconds.
Pagination not documented. Tools returning lists (list_my_calls, list_my_notes, list_team, team_activity, closer_activity, find_leads, find_my_leads) lack limit/offset parameters and total count in descriptions. Large result sets could exhaust context.
Parameter relationship undocumented. closer_activity accepts 'closer' as union(string|number) but descriptions do not clarify when to use name vs ID, or what happens if both are ambiguous.
Manager-only access control not enforced in tool definitions. Tools like list_team, team_activity, team_pipeline, find_leads, lead_activity declare 'Manager only' in descriptions but lack explicit permission annotations or schema constraints.