Loading ToolBench

Reading the current corpus…

hob_hunt_mcp MCP server · D · ToolBench

← Browse

hob_hunt_mcp

github.com/Corykidios/hob_hunt_mcp

A unified MCP server for web search and page fetching — hob_search, hob_site, hob_sites.

2026-07-28+STDIO2 tools

54/100

Definition Quality72

Protocol Readiness50

Supportability26

Spec posture

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

Static source inference · medium confidence

Current-spec patterns detected

Summary

Two tools with comprehensive parameter schemas and detailed descriptions. hunt_site has 14 well-typed parameters with clear descriptions; hunt_search has 6 parameters. Both tools document their purpose and use cases. However, output schemas are not explicitly documented in the source code, only input schemas are visible. Error handling is present but generic (no recovery guidance or categorization). Tool names lack action verbs (hunt_site, hunt_search are descriptive but not verb-noun pattern). No tool annotations (readOnlyHint, destructiveHint) despite both being read-only operations.

Tools (2)

hunt_searchread onlysource verified70/100

Perform Google searches with anti-bot stealth, CAPTCHA handling, and result extraction. Supports multiple queries, result limits, locale/timezone spoofing, and session state persistence.

hunt_siteread onlysource verified73/100

Fetch and extract content from one or more URLs. Supports single-page fetch, multi-page crawl, CSS selector targeting, content extraction via Readability, HTML/Markdown output, link extraction, and sandbox wrapping for security.

Issues (5)

MEDIUMtool-naming ↗

Tool names lack action verbs. 'hunt_site' and 'hunt_search' are descriptive but do not follow verb_noun pattern (e.g., fetch_page, search_web). LLMs infer intent from verb-first names; these names require reading descriptions to disambiguate.

hunt_sitehunt_search

HIGHresponse-shaper ↗

Output schemas not documented in source code. Only input schemas are visible. LLMs cannot plan downstream calls or extract fields without knowing what hunt_site and hunt_search return. Baseline: 100% of A+ tools document return types.

hunt_sitehunt_search

MEDIUMtool-annotations ↗

No tool annotations despite both tools being read-only. Missing readOnlyHint annotation prevents clients from optimizing caching and retry logic. Current spec (2026-07-28) rewards tool annotations.

hunt_sitehunt_search

HIGHrecovery-guide ↗

Recommendations

Rename tools to verb_noun pattern: hunt_site → fetch_page or fetch_content; hunt_search → search_web or search_google. This aligns with 90% of A+ tools and improves LLM intent inference.
Document output schemas explicitly in tool descriptions or code comments. Specify fields returned by hunt_site (e.g., content, links, metadata) and hunt_search (e.g., results array with title/link/snippet). Include example response structure.
Add readOnlyHint: true annotation to both tools in tool registration. This signals to clients that these tools are safe for caching and retry without side effects.
Enhance error handling with recovery guidance. Return structured errors like {error: 'URL_TIMEOUT', message: 'Page load exceeded 30s. Try increasing timeout or using hunt_search to find a cached version.', retryable: true}.
Add validation constraints to hunt_site 'pages' parameter description: 'Number of pages to crawl (1-10, BFS, same-domain only; default 1). Higher values may trigger rate limits or timeouts.'

Score history

Overall score trend

First recorded score · v2 rubric

54/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	D	54	2026-07-28+	v2