Loading ToolBench
High-performance, local-first website crawler and technical SEO audit engine with MCP integration for AI agent pair-programming
Black Sparrow demonstrates solid tool design with complete JSON schemas, clear descriptions, and well-structured parameters. All 8 tools have explicit definitions with proper type constraints (enums, min/max bounds). Descriptions average ~150 chars and explain WHAT, WHEN, and WHY. However, output schemas are not documented, LLMs cannot predict response structure. Error handling is minimal; tools lack recovery guidance. Parameter descriptions are good but could be more prescriptive about constraints. No tool annotations (readOnlyHint/destructiveHint) despite clear risk profiles.
Polls the live progress and operational telemetry of an active or finished crawl session.
Audits whether a website is optimized for Generative Engine Optimization (GEO) and AI Search Engines (ChatGPT Search, Perplexity, Claude) and checks /llms.txt.
Deletes audit records and artifacts from disk for a crawl session.
Generates a structured, token-efficient Markdown audit report engineered specifically for LLM context windows (zero ANSI codes, zero ASCII art, high signal-to-noise ratio).
Queries specific issues from an audit database, filtered by severity, category, or URL pattern.
Performs an instant, synchronous audit of a single URL in < 500ms. Ideal for testing a specific landing page or verifying a code fix immediately.
Output schemas not documented. LLMs cannot predict response structure for any tool, forcing them to guess what fields are returned and breaking downstream tool chaining.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). seo_cleanup_session is destructive but unmarked; seo_start_audit is idempotent but unmarked. LLMs cannot reason about retry safety or side effects.
Error handling lacks recovery guidance. Tools return errors but do not tell LLMs what to do next (retry, ask user, call alternative tool). E.g., 'session not found' should suggest 'Call seo_start_audit to create a new session.'
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 61 | <=2025-11-25 | v2 |
Kicks off a background crawl and technical SEO audit for an entire website. Returns immediately with a session_id in < 1.0s.
Validates a raw JSON-LD or Schema.org block against Google Rich Results eligibility rules.
seo_validate_schema parameter 'json_ld' accepts string but lacks format/pattern constraint. Description says 'raw JSON-LD string or object snippet' but schema only allows string type, ambiguous for LLM.
seo_query_issues 'severity' and 'category' parameters are optional but lack defaults. LLM must decide whether to omit them or guess values, risking unintended filtering.