catalan-adobe/browser-universal
Detect the available browser interaction layer and load the right commands — then navigate, click, fill, and screenshot through a unified verb set. playwright-cli is the default, recommended layer; falls back to Playwright MCP, cmux-browser, or CDP when it is absent. Use before any browser interaction in skills that shouldn't hardcode a specific layer. Triggers on: browser universal, detect browser, browser layer, browser setup, which browser, browser interaction, open browser, use browser.
77
97%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong, well-crafted description that clearly communicates its purpose as a browser interaction abstraction layer. It specifies concrete actions, lists the technology fallback chain, provides explicit 'when to use' guidance, and includes comprehensive trigger terms. The description is concise yet thorough, making it easy for Claude to distinguish this skill from other browser-related skills.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple concrete actions: 'detect the available browser interaction layer', 'load the right commands', 'navigate, click, fill, and screenshot'. Also specifies the fallback chain: playwright-cli, Playwright MCP, cmux-browser, CDP. | 3 / 3 |
Completeness | Clearly answers both 'what' (detect browser layer, load commands, provide unified verb set for navigation/clicking/filling/screenshots) and 'when' ('Use before any browser interaction in skills that shouldn't hardcode a specific layer') with explicit trigger terms listed. | 3 / 3 |
Trigger Term Quality | Includes a comprehensive set of natural trigger terms: 'browser universal', 'detect browser', 'browser layer', 'browser setup', 'which browser', 'browser interaction', 'open browser', 'use browser'. These cover many natural ways a user or another skill might reference browser interaction. | 3 / 3 |
Distinctiveness Conflict Risk | Occupies a clear niche as a browser abstraction/detection layer rather than a specific browser automation skill. The focus on detecting and selecting the right browser interaction layer is distinct from skills that simply perform browser automation. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
92%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-crafted skill that efficiently handles a complex multi-layer detection and abstraction problem. The detection ladder is clear and actionable with executable commands, the universal verbs table is an excellent compact reference, and the workflow includes proper validation patterns. The only weakness is that the referenced LAYERS.md bundle file isn't provided, making it impossible to verify the progressive disclosure chain is complete.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is lean and efficient. It doesn't explain what a browser is or how detection works conceptually — it jumps straight into the detection ladder with concrete commands. Every section earns its place, and the universal verbs table is a compact, high-density reference. | 3 / 3 |
Actionability | Each detection step has executable shell commands with clear success criteria (exit codes, tool existence checks). The universal verbs table provides exact command mappings. The error case includes specific remediation steps. The targeting model distinction (ref-based vs selector-based) is concrete and actionable. | 3 / 3 |
Workflow Clarity | The detection ladder is explicitly sequenced with a clear 'first one wins, stop checking' rule. The universal pattern section provides a critical validation checkpoint (re-snapshot after state changes). The 'No Layer Detected' section is a clear blocking error with recovery guidance. The workflow from detection → load reference → use verbs is well-sequenced. | 3 / 3 |
Progressive Disclosure | The skill references `references/LAYERS.md` for detailed layer-specific commands, which is good progressive disclosure. However, no bundle files were provided, so we can't verify the reference exists. The universal verbs table is appropriately inline as a quick reference, but the skill is somewhat long and could potentially split more detail into the referenced file. | 2 / 3 |
Total | 11 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
Reviewed
Table of Contents