browser-sessions

Architecture for attaching Chrome page context and control to Agent-Native conversations. Use when building browser capture, extension chat, browser control, or cross-app webpage workflows.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

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 lean, highly actionable architecture spec that names concrete APIs and primitives while weaving approval/validation gates into the control flow. It is well-sectioned and token-efficient; the main gap is the absence of an explicit numbered workflow with error-recovery feedback loops and any one-level-deep reference structure.

Suggestions

Add a short numbered control-lease workflow (attach → observe → act → stop) with an explicit validate/fix/retry loop for failed actions to push workflow_clarity to 5.

Consider splitting the per-action relay reference table and the BrowserContextV1 projection details into a single one-level-deep reference file so SKILL.md stays a tighter overview.

Include one minimal executable snippet (e.g. a scripting.executeScript capture call or a bridge-message validation guard) to make the guidance copy-paste concrete.

Dimension	Reasoning	Score
Conciseness	The body is a dense set of architectural constraints that assumes Claude's competence ('Read capture never attaches the debugger', 'Do not capture hidden content, form values, password fields...') with no padded explanations of basic concepts; nearly every token carries a rule.	5 / 5
Actionability	Although code-free, this instruction-only skill names exact primitives (`BrowserContextV1`, `parseBrowserContextV1`, `@agent-native/core/browser-context`, `chrome.debugger`, `scripting.executeScript`, the four relay actions) and gives crisp do/don't directives, which per the scoring note is actionable without penalizing the absence of code.	5 / 5
Workflow Clarity	The attach→lease→observe→act flow is clear ('attach establishes the control lease; subsequent target handles must come from the latest observe') with explicit validation boundaries (inline approval gate, 'validate the action type... before enqueueing and again at the executor boundary'), but it is not presented as a numbered checklist with error-recovery feedback loops.	4 / 5
Progressive Disclosure	The body is well-organized into clearly labeled sections (Capability tiers, Context artifact, Side-panel chat, Control transports, Agent tool surface, Product workflows) with no external bundle files to navigate, but as a single ~120-line dense spec it stays monolithic where some detail could plausibly be split out.	4 / 5
	Total	18 / 20 Passed

Description

87%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 pairs a clear capability statement with an explicit 'Use when' trigger clause covering several concrete browser-work scenarios. It is highly distinct and complete, with only minor room for more granular action wording and synonym coverage.

Dimension	Reasoning	Score
Specificity	Names the domain ('attaching Chrome page context and control to Agent-Native conversations') and lists several concrete capability areas ('browser capture, extension chat, browser control, or cross-app webpage workflows'), but these read as workflow categories rather than fully concrete actions like 'extract text, fill forms'.	4 / 5
Completeness	It explicitly answers both what ('Architecture for attaching Chrome page context and control...') and when ('Use when building browser capture, extension chat, browser control, or cross-app webpage workflows') with concrete trigger phrases.	5 / 5
Trigger Term Quality	'browser capture, extension chat, browser control, or cross-app webpage workflows' are natural developer-facing triggers with good coverage, but it lacks synonyms and the specialized domain has no file-extension-style markers the 5 anchor calls for.	4 / 5
Distinctiveness Conflict Risk	It carves a clear niche (Chrome page context/control bound to Agent-Native conversations) with distinct triggers and minimal overlap risk against other skills.	5 / 5
	Total	18 / 20 Passed

Validation

81%

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 — 13 / 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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: BuilderIO/agent-native
Path: .agents/skills/browser-sessions/SKILL.md
Commit: b0df876

Reviewed: 5 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.