MCP (Model Context Protocol) server for browser tools integration, providing access to browser logs, network data, screenshots, and Lighthouse audits
Critical gaps across all dimensions. Descriptions are extremely brief (10-30 chars), below the 20-char floor for scoring above 0-20. No parameter descriptions exist. No output schemas documented. No error handling guidance. Tools lack specificity: 'Check our browser logs' vs 'Check our browsers console errors' are nearly identical. No tool annotations (readOnlyHint/destructiveHint). The server is STDIO-only, which caps protocol readiness at 50 and signals non-production readiness.
Check our browsers console errors
Check our browser logs
Check our network ERROR logs
Check ALL our network logs
Get the selected element from the browser
Run an accessibility audit on the current page
Take a screenshot of the current browser tab
All tools have empty input schemas ({}). This violates pattern:constrained-input and pattern:tool-description.
Descriptions are 10-30 characters, below the 20-char floor. Examples: 'Check our browser logs' (22 chars), 'Check our browsers console errors' (33 chars). Descriptions must explain WHAT, WHEN, and any prerequisites.
No output schemas documented. LLMs cannot plan downstream calls or extract data without knowing return structure. Violates pattern:tool and pattern:response-shaper.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 34 | 2026-07-28+ | v2 |
Wipe all browser logs from memory
No error handling guidance. Tools do not indicate what to do on failure (retryable, user-fixable, fatal). Violates pattern:recovery-guide and pattern:error-classification.
Destructive tool 'wipeLogs' lacks confirmation/dry-run pattern. No description states it modifies state. Violates pattern:confirmation-request and pattern:command-tool.