opencli-browser

Use when an agent needs to drive a real Chrome window via opencli — inspect a page, fill forms, click through logged-in flows, or extract data ad-hoc. Covers the selector-first target contract, compound form fields, stale-ref handling, network capture, and the agent-native envelopes the CLI returns. Not for writing adapters — see opencli-adapter-author for that.

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Quality

Content

81%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 highly actionable, well-sequenced browser-driving skill with executable recipes, explicit validation checkpoints, and clear error-handling guidance. Its main weaknesses are length/repetition (conciseness) and the absence of any reference files to offload the large command-reference and envelope-detail sections (progressive disclosure).

Suggestions

Move the full Command reference and Structured error codes tables into a references/ file (e.g. REFERENCE.md) and keep only the most-used commands inline, reducing repetition of match_level and ref-handling guidance across Mental model, Target contract, Critical rules, and Pitfalls.

Consolidate the match_level explanation (currently restated in Mental model, the match_level table, Critical rule 4, and Pitfalls) into one authoritative section to trim tokens without losing signal.

The Network and Extract subsections each carry verbose flag prose; consider a compact table-plus-one-example form similar to the Wait section to tighten them.

Dimension	Reasoning	Score
Conciseness	The body is dense and largely earns its tokens with concrete command tables, envelope shapes, and cost guidance, but at ~440 lines it is well above a lean overview and repeats several explanations (e.g., match_level and re-identified refs are explained in 'Mental model', the 'Target contract' table, 'Critical rules', and 'Pitfalls'), which keeps it at 'mostly efficient but could be tightened'.	3 / 5
Actionability	Nearly every section gives copy-paste-ready bash commands and concrete JSON envelope examples covering the common cases (login form fill, long dropdown, React dropdown, network scrape, chunked extract, cross-origin iframe); specific flags and return shapes are documented throughout.	5 / 5
Workflow Clarity	Multi-step flows are explicitly sequenced with validation checkpoints — 'Always inspect before you act', 'state -> action -> state after a page change', 'Verify writes that matter' with get value, and the recipes show full state->type->verify->click->wait->state sequences with explicit feedback loops for stale refs and autocomplete.	5 / 5
Progressive Disclosure	Well-organized with clear section headers and a useful 'See also' pointer to sibling skills, but there are no bundle files in references/scripts/assets — all content is inlined in SKILL.md, so it is structured but does not demonstrate the one-level-deep file split that defines a 5.	4 / 5
	Total	17 / 20 Passed

Description

83%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, specific description that clearly states what the skill does and when to use it, with good trigger phrases and explicit de-confliction against a sibling skill. Slightly short of fully comprehensive keyword coverage.

Dimension	Reasoning	Score
Specificity	Names several concrete actions (inspect a page, fill forms, click through logged-in flows, extract data) plus specific capability areas (selector-first target contract, compound form fields, stale-ref handling, network capture); a couple of actions are phrased somewhat generically, so it sits just below a fully comprehensive 5.	4 / 5
Completeness	Explicitly answers both 'what' (drive a Chrome window via opencli; inspect, fill forms, click flows, extract data; covers specific contracts) and 'when' ('Use when an agent needs to drive a real Chrome window via opencli ...') with concrete trigger phrases.	5 / 5
Trigger Term Quality	Includes natural user-facing phrases like 'drive a real Chrome window', 'fill forms', 'click through logged-in flows', and 'extract data', but lacks synonyms and the tool/file-extension variants that would push it to 5.	4 / 5
Distinctiveness Conflict Risk	Has a clear niche (driving a live browser via opencli) and explicitly redirects overlapping work to a sibling skill ('Not for writing adapters — see opencli-adapter-author for that'), minimizing conflict risk.	5 / 5
	Total	18 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning

	Total	15 / 16 Passed

Repository: jackwener/OpenCLI
Path: skills/opencli-browser/SKILL.md
Commit: 64e6f0e

Reviewed: 2 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.