Loading ToolBench

Reading the current corpus…

Hi-LLM MCP server · F · ToolBench

Hi-LLM

HiLLM NAV is an open-source navigation and community platform for AI tools, Skills, MCP servers, and Prompts. Includes security scanning and assessment infrastructure.

2026-07-28+STREAMABLE_HTTP5 tools

30/100

Definition Quality24

Protocol Readiness35

Supportability38

Spec posture

Inferred effective spec 2026-07-28+·current baseline 2026-07-28

Static source inference · medium confidence · evidence: Streamable HTTP

Current-spec patterns detected

Summary

This server exhibits severe definition quality gaps across all dimensions. Tool names lack action verbs (scan_skill, chat_completions, health_live, health_ready, ensureUniqueSkillSlug are inconsistent and vague). Descriptions are present but minimal (10-50 chars for most tools). Input schemas are visible but lack proper type constraints and validation guidance. No output schemas documented. Error handling is absent from descriptions. The server appears to be a Next.js application exposing internal services rather than a proper MCP tool server, tool definitions are inferred from file paths and brief descriptions, not explicitly registered with full schemas.

Tools (5)

chat_completionsread onlyauth25/100

Gateway endpoint for LLM chat completions used by the security scanner for AI-powered vulnerability assessment

ensureUniqueSkillSlugread only38/100

Generates a unique URL slug for a Skill, avoiding collisions with existing skills

health_liveread only27/100

Liveness probe endpoint for container orchestration

health_readyread only28/100

Readiness probe endpoint that verifies scanner configuration and model availability

scan_skillread onlyauth33/100

Scans a Skill package (ZIP archive) for security vulnerabilities and issues using aig-skill-scan

Issues (5)

CRITICALtool-naming ↗

Tool names lack consistent action verbs. 'scan_skill' is acceptable, but 'chat_completions', 'health_live', 'health_ready', and 'ensureUniqueSkillSlug' violate verb_noun convention. LLMs cannot infer intent from these names alone.

chat_completionshealth_livehealth_readyensureUniqueSkillSlug

CRITICALresponse-shaper ↗

No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields. Responses are unstructured.

scan_skillchat_completionshealth_livehealth_readyensureUniqueSkillSlug

HIGHtool-description ↗

Descriptions are too brief (10-50 chars). Baseline for A+ tools is 50-200 chars. Current descriptions lack context for when/why to use each tool. E.g., 'Liveness probe endpoint' does not explain what an LLM should do with it.

health_livehealth_readychat_completions

Recommendations

Rename tools to follow verb_noun convention: 'chat_completions' → 'get_chat_completion', 'health_live' → 'check_liveness', 'health_ready' → 'check_readiness', 'ensureUniqueSkillSlug' → 'generate_unique_skill_slug'.
Expand all descriptions to 50-200 characters. Include WHAT the tool does, WHEN to use it, and what it returns. E.g., 'scan_skill: Analyzes a ZIP-packaged Skill for security vulnerabilities using aig-skill-scan. Returns a report with vulnerability count, severity levels, and remediation guidance. Use before publishing a Skill.'
Document output schemas for all tools. Specify field names, types, and what data each field contains. E.g., scan_skill should return {vulnerabilities: [{id, severity, description, remediation}], total_count, scan_duration_ms}.
Add error handling guidance to descriptions. E.g., 'scan_skill: Returns 400 if ZIP is malformed (try re-packaging). Returns 503 if scanner is unavailable (retry after 30s). Returns 422 if SKILL.md is missing (required in archive root).'
Verify tool definitions are explicitly registered in MCP server code with full input/output schemas. If tools are inferred from HTTP endpoints, refactor to use MCP's tool registration API with complete schemas.

Score history

Overall score trend

First recorded score · v2 rubric

30/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	30	2026-07-28+	v2