Loading ToolBench
Give your AI agent read/search access to no-login, low-risk web sources.
NetReach provides 5 tools with basic schemas and descriptions, but falls short of production quality. All tools have descriptions (positive), but they are brief (avg ~80 chars, below the 194-char baseline) and lack actionable context. Schemas are present but minimal: most parameters have type and description, but lack constraints (enums, ranges, patterns). No output schemas documented. Error handling is absent, no guidance on retryability, recovery steps, or invalid input handling. Tool names follow verb_noun convention (get_*, read) which is good, but descriptions are generic and do not explain WHEN to use each tool or what distinguishes them. No tool annotations (readOnlyHint, idempotentHint) despite all being read-only. Parameter descriptions are terse (e.g., 'The URL to read' for read tool, 'Maximum number of topics to return' for get_hot_topics) and lack format guidance or constraints.
获取热门帖子列表。Returns a list of dicts with keys: title, url, replies, node_name, node_title, content
获取指定节点的最新帖子。Returns a list of dicts with keys: title, url, replies, node_name, node_title, content
获取单个帖子详情和回复列表。Returns a dict with keys: id, title, url, content, replies_count, node_name, node_title, author, created, replies (list of dicts with: author, content, created)
获取用户信息。Returns a dict with keys: id, username, url, website, twitter, psn, github, btc, location, bio, avatar, created
通过 Jina Reader 读取网页，返回 Markdown 全文。
No output schemas documented. Tool descriptions mention return fields (e.g., 'Returns a list of dicts with keys: title, url, replies...') but no formal schema. LLMs cannot plan downstream calls or extract fields reliably without structured output documentation.
Descriptions are too brief (avg 80 chars vs 194-char baseline) and lack actionable context. No explanation of WHEN to use each tool, dependencies between tools, or how they differ. E.g., get_hot_topics vs get_node_topics distinction is unclear from descriptions alone.
No error handling guidance. Tools lack descriptions of failure modes, retryability, or recovery steps. E.g., what happens if a topic_id is invalid? Should the LLM retry, search for the topic, or ask the user?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 46 | 2026-07-28+ | v2 |
Parameter descriptions lack constraints and format guidance. 'limit' parameter has no min/max bounds (baseline: 1 - 100 or similar). 'url' parameter has no format hint (http/https only?). 'node_name' has no enum or pattern (e.g., 'python', 'tech', 'jobs' are examples but not constraints).
No tool annotations despite all tools being read-only. Missing readOnlyHint annotation would signal to agents that these tools are safe to call without confirmation, improving planning efficiency.