Loading ToolBench
MCP server for email security auditing with DMARC, SPF, and DKIM validation to prevent spoofing and impersonation attacks
Single tool 'execute_skill_action' has critical definition gaps. Input schema present but minimal (only 'domain' string with generic default). Description is adequate (67 chars) but lacks WHEN/WHY guidance. No output schema documented anywhere, LLM cannot plan downstream actions. No error handling, no parameter validation hints, no idempotency guidance. Tool name uses 'execute' (generic verb) rather than action-specific verb like 'audit_domain_security'. Implementation is mock-only (hardcoded responses), not production-ready. No security considerations documented despite handling email infrastructure data.
Execute email security audit and DMARC spoofing defense analysis
No output schema documented. LLM cannot infer response structure (fields, types, nesting). Prevents downstream tool chaining and forces LLM to guess field names.
Tool name 'execute_skill_action' is generic and vague. Does not start with action verb reflecting actual behavior (audit, analyze, validate). LLM cannot infer intent from name alone.
Input parameter 'domain' lacks validation constraints. No format specification, no example of valid domain format, no guidance on what happens with invalid input. Description does not state expected format or constraints.
No error handling documented. Tool description does not explain failure modes, recovery steps, or what errors the LLM should expect. No actionable error messages in implementation.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 38 | 2026-07-28+ | v2 |
Implementation is mock-only with hardcoded responses. No actual DNS lookups, DMARC validation, or security analysis. Tool cannot be used for real email security audits.