Loading ToolBench
MCP server providing data and tools grounded in Engineering Leaders Community's 3,300+ engineering leaders. Offers leadership-ratio benchmarking, membership business-case building, and community-launch readiness assessment.
ELC Toolkit demonstrates solid definition quality with complete schemas, clear descriptions, and thoughtful parameter design. All 5 tools have explicit registrations with descriptions (194 - 392 chars, within baseline), input schemas with typed parameters, and output schemas. Tool names follow verb_noun convention (get_*, benchmark_*, build_*, assess_*). Annotations are present and correct (READ_ONLY contract). However, some parameter descriptions lack actionable constraints (e.g., 'context' in get_more_tools is vague about expected format), and error handling is minimal, no recovery guidance or categorization visible in the code. Output is well-structured (report, source, verdict fields) but lacks pagination/limit guidance for potential future expansion.
Community launch readiness test — should you start a local engineering-leadership meetup in your city? Built directly from ELC's own new-city playbook, the real operating doc used to launch Brno, Bratislava and Kraków. Not theory — the actual checklist. Checks five dimensions: speaker network, promotion channels, local operator, realistic targets, and promotion budget.
Describes a company's manager-vs-senior-IC split — each side's percentage, the resulting span (1 manager per N senior ICs), and the question that shape usually raises. Alongside it, for context rather than as a target, ELC's own community composition (69% Manager+/Leadership, 21% Senior/Staff IC across 3,300+ CEE engineering leaders — that is who joins a leadership community, not a survey of org structures, so there is deliberately no score against it). Count the SAME population on both sides: senior people who could plausibly hold a management role (managers, tech leads, senior/staff ICs), leaving out junior/mid ICs.
Builds the internal business case for partnering with Engineering Leaders Community: real reach (3,300+ engineering leaders), membership tiers, pricing, and the approval email template to send up the chain.
Check for additional tools whenever your task might benefit from specialized capabilities, even if existing tools could work as a fallback. Also the right tool for a bare greeting (hi, hello), a connectivity/liveness test, or any message too general to match a specific tool below — pass it as `context`, or send no arguments at all, and this returns the full menu instead of a dead end.
Parameter 'context' in get_more_tools lacks actionable format guidance. Description says 'A description of your goal' but does not specify length, structure, or examples of valid inputs. LLMs may pass ambiguous or overly long context.
No error handling or recovery guidance visible in tool implementations. If invalid enum values are passed (e.g., goal='invalid' in build_partnership_business_case), the code does not show how errors are formatted or what the LLM should do next.
assess_community_readiness accepts 'answers' as a record with arbitrary string keys, but only validates against READINESS_DIMENSIONS at runtime. The schema does not enforce enum keys, allowing invalid dimension IDs to be passed without early rejection.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | B | 73 | 2026-07-28+ | v2 |
Call this for a greeting (hi, hello), a connectivity/liveness test, 'what can you do', or any message too general to match a specific tool below. Returns the full menu of real questions this server answers, each mapped to the tool name that answers it, so the next call can go straight to the right tool.
Output schema (REPORT_OUTPUT) is shared across all tools but 'verdict' is optional and only some tools populate it. The description does not clarify which tools return a verdict, forcing LLMs to handle inconsistent response shapes.
No pagination or result-limiting guidance in tool descriptions. If tools are extended to return lists (e.g., list_partnership_tiers), there is no established pattern for limit/offset or cursor-based pagination.