Loading ToolBench

Reading the current corpus…

callrail-mcp MCP server · C · ToolBench

← Browse

callrail-mcp

github.com/pghdma/callrail-mcp

Model Context Protocol (MCP) server for the CallRail REST API v3

<=2025-11-25STDIO15 tools

61/100

Definition Quality79

Protocol Readiness42

Supportability43

Spec posture

Inferred effective spec <=2025-11-25·current baseline 2026-07-28

Static source inference · medium confidence · detected: Logging

Deprecated protocol patterns detected

Summary

CallRail MCP has well-structured tool definitions with consistent naming (verb_noun pattern), clear descriptions (avg 80-120 chars), and complete JSON schemas for all 15 tools. All parameters are typed and described. However, output schemas are not documented in the source code, error handling lacks recovery guidance, and some parameter descriptions could be more prescriptive about constraints. No tool annotations (readOnlyHint/destructiveHint) despite clear risk classifications. Composition is sound, tools are single-responsibility and chainable via IDs.

Tools (15)

compare_periodsread onlyauthsource verified73/100

Compare call metrics between two date periods

create_tagwriteauthsource verified75/100

Create a new tag with specified name and color

create_trackerwriteauthsource verified77/100

Create a new call tracker

delete_trackerdestructiveauthsource verified77/100

Delete a call tracker

get_callread onlyauthsource verified83/100

Get details of a specific call by ID

get_companyread onlyauthsource verified83/100

Get details of a specific company by ID

get_trackerread onlyauthsource verified83/100

Get details of a specific tracker by ID

Issues (5)

HIGHtool ↗

Output schemas not documented. Tool descriptions state what is returned (e.g., 'Get details of a specific call') but do not specify the response structure, field names, or types. LLMs cannot plan downstream tool calls or extract specific fields without knowing the response schema.

list_callsget_calllist_companiesget_companylist_tagscreate_tagtag_calluntag_calllist_trackersget_trackercreate_trackerupdate_trackerdelete_trackerusage_summarycompare_periods

HIGHtool-annotations ↗

Missing tool annotations. Tools are classified by risk (READ_ONLY, WRITE, DESTRUCTIVE) but do not use MCP tool annotations (readOnlyHint, destructiveHint, idempotentHint). This prevents clients from applying appropriate safety guardrails or UI affordances.

delete_trackercreate_tagtag_calluntag_callcreate_trackerupdate_tracker

Recommendations

Document output schemas for all tools. Add a 'Returns' section to each tool description specifying field names, types, and structure. Example: 'Returns: {call_id: string, duration_seconds: integer, caller_phone: string, tags: [string], created_at: ISO8601 timestamp}'.
Add tool annotations to the MCP tool registration. Use readOnlyHint=true for all READ_ONLY tools, destructiveHint=true for delete_tracker, and idempotentHint=true for get_* and list_* tools. This enables client-side safety guardrails.
Enhance error messages with recovery guidance. When tag_call fails because a tag does not exist, return: 'Tag "urgent" not found. Available tags: [list]. Call list_tags() to see all tags.' When create_tracker fails due to invalid phone_number, return: 'Invalid phone_number format. Expected E.164 format (e.g., +1-555-0123).'
Specify parameter constraints inline in descriptions. For create_tag, change 'Tag color (must be one of the valid CallRail colors)' to 'Tag color. Must be one of: red, blue, green, yellow, orange, purple, pink, gray.' For phone_number, specify: 'Phone number in E.164 format (e.g., +1-555-0123) or domestic format (555-0123).'
Document pagination behavior. For list_calls and list_trackers, add: 'Pagination: page parameter (1-indexed). Results capped at 50 per page. Returns total_count and has_more fields to enable iteration.' Specify the default page size and maximum.

Spec posture evidence

Inferred effective spec: <=2025-11-25.

Relies on Logging (deprecated) - log to stderr or use OpenTelemetry

Score history

Overall score trend

First recorded score · v2 rubric

61/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	C	61	<=2025-11-25	v2