chrome-devtools-cli
Use this skill to write shell scripts or run shell commands to automate tasks in the browser or otherwise use Chrome DevTools via CLI.
68
57%
Does it follow best practices?
Impact
81%
1.22xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/chrome-devtools-cli/SKILL.mdQuality
Discovery
50%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
The description provides a reasonable overview of the skill's domain—bridging shell scripting with Chrome DevTools browser automation—but lacks specific concrete actions and explicit trigger guidance. It uses second person voice ('Use this skill') which slightly undermines professionalism, and the trigger terms could be expanded to cover more natural user phrasings.
Suggestions
Add explicit 'Use when...' clause with natural trigger terms like 'browser automation', 'headless Chrome', 'CDP', 'DevTools protocol', 'web scraping via CLI'.
List specific concrete actions such as 'capture screenshots', 'extract page content', 'monitor network requests', 'run headless browser sessions' to improve specificity.
Rewrite in third person voice (e.g., 'Writes shell scripts and runs shell commands to automate browser tasks via Chrome DevTools CLI') instead of the imperative 'Use this skill to...'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | It names the domain (shell scripts, Chrome DevTools, browser automation) and some actions (write shell scripts, run shell commands, automate tasks), but doesn't list specific concrete actions like 'capture screenshots', 'inspect network requests', or 'extract DOM elements'. | 2 / 3 |
Completeness | The 'what' is partially addressed (write shell scripts, run commands for browser automation via Chrome DevTools CLI), but the 'when' is only implied through 'Use this skill to...' rather than providing explicit trigger conditions like 'Use when the user asks about browser automation, Chrome DevTools, or running headless browser commands'. | 2 / 3 |
Trigger Term Quality | Includes some relevant keywords like 'shell scripts', 'shell commands', 'Chrome DevTools', 'CLI', and 'browser', but misses common user variations like 'devtools protocol', 'headless Chrome', 'browser automation', 'CDP', 'puppeteer', or 'web scraping'. | 2 / 3 |
Distinctiveness Conflict Risk | The combination of shell scripting and Chrome DevTools via CLI is somewhat distinctive, but 'shell scripts' and 'automate tasks' are broad enough to overlap with general shell scripting or automation skills. The Chrome DevTools angle provides some differentiation but isn't strongly emphasized. | 2 / 3 |
Total | 8 / 12 Passed |
Implementation
64%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill excels at actionability with concrete, executable commands and a clear snapshot example showing the UID-based interaction model. However, it is overly verbose by exhaustively listing every command variant inline rather than using progressive disclosure to a reference file. The workflow section is adequate but lacks validation checkpoints for stateful browser automation sequences.
Suggestions
Move the exhaustive command listings into a separate reference file (e.g., references/commands.md) and keep only the most common 2-3 examples per category in SKILL.md.
Add validation/verification guidance to the workflow, e.g., 'After click/fill, take_snapshot to verify the expected state change before proceeding.'
Condense command examples by showing the base pattern once per category and noting that --help and --includeSnapshot are available on most commands, rather than repeating variants.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient and avoids explaining concepts Claude already knows, but the exhaustive listing of every command variant with comments is verbose. Many of these could be condensed (e.g., showing the pattern once and noting flags exist via --help) rather than enumerating every permutation. | 2 / 3 |
Actionability | Every command is concrete, copy-paste ready, and includes real arguments and flags. The snapshot example shows actual output format with UIDs, making it immediately clear how to chain commands together. | 3 / 3 |
Workflow Clarity | The 3-step AI Workflow (Execute → Inspect → Act) provides a clear high-level sequence, but there are no validation checkpoints or error recovery steps. For browser automation involving stateful interactions, missing feedback loops (e.g., verifying a click succeeded via snapshot before proceeding) is a gap. | 2 / 3 |
Progressive Disclosure | The skill references installation.md for setup (good), but the massive command reference is entirely inline rather than split into a separate reference file. The content would benefit from keeping only the core workflow and most common commands in SKILL.md, with the full command catalog in a referenced file. | 2 / 3 |
Total | 9 / 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.
- Repository
- ChromeDevTools/chrome-devtools-mcp
- Commit
2e039c0
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.