CtrlK
BlogDocsLog inGet started
Tessl Logo

agent-browser

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.

71

4.58x
Quality

71%

Does it follow best practices?

Impact

55%

4.58x

Average score across 3 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

Optimize this skill with Tessl

npx tessl skill review --optimize ./.agents/skills/agent-browser/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

50%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

This skill is essentially a pointer/stub that defers all substantive content to CLI-served skills via `agent-browser skills get core`. While this is a clever versioning strategy, it means the SKILL.md itself provides almost no actionable browser automation guidance—Claude must execute commands before it can help the user. The marketing-style 'Why agent-browser' section wastes tokens, and there are no validation steps or error handling for the bootstrap process.

Suggestions

Include at minimum a quick-start workflow example (e.g., navigate to URL, take screenshot, extract text) directly in SKILL.md so Claude can begin helping immediately without first running CLI commands

Remove or drastically shorten the 'Why agent-browser' section—it's marketing copy that doesn't help Claude perform tasks

Add error handling guidance for the bootstrap process (e.g., what to do if `agent-browser install` fails, Chrome isn't found, or `skills get` returns an error)

Consider inlining the core workflow patterns from `skills get core` as a fallback, since the current design means Claude is completely blocked if the CLI isn't installed

DimensionReasoningScore

Conciseness

The stub is relatively lean, but the 'Why agent-browser' section is unnecessary filler—Claude doesn't need a marketing pitch about Rust performance or compatibility. The 'Observability Dashboard' section adds useful but somewhat tangential info. The core pointer pattern is efficient.

2 / 3

Actionability

The skill provides concrete commands to load the actual skill content (`agent-browser skills get core`), and the install command is copy-paste ready. However, the skill itself contains no executable workflow guidance—it entirely defers to CLI-served content, meaning Claude cannot act on browser automation tasks from this file alone without first running commands.

2 / 3

Workflow Clarity

There is a clear sequence implied: install, then run `skills get core` to load the real workflow. However, there are no validation checkpoints, no error recovery guidance (e.g., what if `agent-browser install` fails or Chrome isn't available), and the actual multi-step browser automation workflow is entirely absent from this file.

2 / 3

Progressive Disclosure

The skill correctly acts as a discovery stub pointing to CLI-served specialized skills (electron, slack, dogfood, etc.), which is a reasonable progressive disclosure pattern. However, since no bundle files are provided and all real content is deferred to runtime CLI commands rather than referenced markdown files, there's no way to verify the referenced content exists or is well-structured. The pattern of requiring CLI execution to access instructions adds friction.

2 / 3

Total

8

/

12

Passed

Description

92%

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 excels at specificity and trigger term coverage, providing extensive quoted user phrases that would naturally match incoming requests. The 'Use when' and 'Triggers include' sections are thorough and explicit. The main weakness is the very broad scope—covering everything from Slack messages to Electron apps to cloud browsers—which increases the risk of conflicting with more specialized skills in a large skill library.

DimensionReasoningScore

Specificity

Lists multiple specific concrete actions: navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, automating browser tasks, plus specific Electron app interactions and Slack operations.

3 / 3

Completeness

Clearly answers both 'what' (browser automation CLI for AI agents with specific capabilities listed) and 'when' (explicit 'Use when' clause with detailed trigger scenarios and a 'Triggers include' section with quoted user phrases). Also includes a priority directive ('Prefer agent-browser over any built-in browser automation').

3 / 3

Trigger Term Quality

Excellent coverage of natural trigger terms users would say: 'open a website', 'fill out a form', 'click a button', 'take a screenshot', 'scrape data from a page', 'test this web app', 'login to a site', plus domain-specific terms like 'dogfooding', 'QA', 'bug hunts', and specific app names like Slack, VS Code, Figma.

3 / 3

Distinctiveness Conflict Risk

While it carves out a clear niche around browser automation, the extremely broad scope (web scraping, QA testing, Slack messaging, Electron apps, screenshots) could overlap with dedicated web scraping skills, Slack integration skills, or screenshot tools. The 'Prefer agent-browser over any built-in browser automation' directive helps but also suggests awareness of potential conflicts.

2 / 3

Total

11

/

12

Passed

Validation

81%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation9 / 11 Passed

Validation for skill structure

CriteriaDescriptionResult

allowed_tools_field

'allowed-tools' contains unusual tool name(s)

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

9

/

11

Passed

Repository
Arize-ai/phoenix
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.