agent-browser

Browser automation via agent-browser CLI. Use when the user needs to navigate websites, fill forms, click buttons, take screenshots, extract data, or test web apps.

Quality

91%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

High

Do not use without reviewing

Quality

Content

90%Weight 40%Scale 1-5

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

A tight, highly actionable CLI reference with executable examples and a clear core workflow plus sensible validation checkpoints. The main room for improvement is splitting the large command catalog into a separate reference file and adding explicit error-recovery guidance for failed interactions.

Suggestions

Move the lengthy Commands catalog into a references/REFERENCE.md, keeping only the most common commands inline in SKILL.md with a clear pointer, to improve progressive disclosure.

Add an error-recovery checkpoint to the Core Workflow (e.g. re-run `snapshot -i` and retry the ref when a click/fill fails or a wait times out) to raise workflow clarity.

Note expected failure modes (element not found, navigation timeout) and the corresponding corrective command so the feedback loop is explicit.

Dimension	Reasoning	Score
Conciseness	The body is a lean command reference with terse inline comments and almost no prose padding (no explanation of what a browser or CDP is), so every token earns its place; matches the level-5 anchor rather than the minor-trimming level-4 anchor.	5 / 5
Actionability	Commands are fully executable and copy-paste ready, with a worked example workflow (open, snapshot, fill, click, wait) covering common cases, matching the level-5 anchor; not a 4 because there are no missing key details in the examples.	5 / 5
Workflow Clarity	The Core Workflow is a clear numbered sequence (open → snapshot -i → interact via refs → re-snapshot) with an explicit upfront availability check (`which agent-browser`) and a re-snapshot checkpoint, but there is no error-recovery feedback loop (e.g. what to do when a command fails), placing it at level-4 rather than 5.	4 / 5
Progressive Disclosure	Content is organized into clearly headed sections (Core Workflow, Commands, Authentication, Iframes, Parallel Sessions, JavaScript Eval, Cleanup) with no nested references, which is good structure; it stays at 4 rather than 5 because the sizable inline Commands block could be split into a reference file and there are no one-level-deep references to offload detail.	4 / 5
	Total	18 / 20 Passed

Description

92%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, third-person description that clearly pairs capability with explicit use-when triggers and a broad set of concrete actions. The only gap is a handful of natural synonyms (scraping, automation, login) that would round out trigger coverage.

Dimension	Reasoning	Score
Specificity	Lists six concrete actions ("navigate websites, fill forms, click buttons, take screenshots, extract data, or test web apps"), giving comprehensive coverage of the browser-automation domain, matching the level-5 anchor; not a 4 because coverage is broad rather than having minor gaps.	5 / 5
Completeness	Explicitly states what ("Browser automation via agent-browser CLI") and when ("Use when the user needs to navigate websites, fill forms...") with concrete trigger phrases, matching the level-5 anchor; not a 4 because the 'when' clause is concrete and specific rather than weakly implied.	5 / 5
Trigger Term Quality	Strong natural terms ("navigate websites", "fill forms", "click buttons", "take screenshots", "extract data", "test web apps") that users would say, but a few common synonyms are absent (e.g. "web scraping", "automate", "log in"), so it sits above the 3 anchor but short of comprehensive level-5.	4 / 5
Distinctiveness Conflict Risk	"Browser automation via agent-browser CLI" carves a clear niche with distinct triggers (website navigation, form filling, screenshots) and minimal overlap risk with other skills; fits the level-5 anchor, not a 4 because the triggers are clearly distinct rather than having minor overlap.	5 / 5
	Total	19 / 20 Passed

Validation

87%

Warnings & errors only

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

Validation — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning

	Total	14 / 16 Passed

Repository: sipeed/picoclaw
Path: workspace/skills/agent-browser/SKILL.md
Commit: bbf6893

Reviewed: 17 days ago

Table of Contents

Discovery Implementation Validation

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.