claude-computer-use-advanced

Advanced computer use patterns for UI automation, application control, and multi-step workflows using Claude's computer use tool. Use when automating desktop tasks, testing applications, analyzing screen content, controlling software programmatically, or building computer vision workflows. Supports zoom tool for enhanced vision on Opus 4.5, multi-step automation, and sophisticated application control.

Quality

75%

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

Fix and improve this skill with Tessl

tessl review fix ./skills/claude-computer-use-advanced/SKILL.md

Quality

Content

67%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 thorough, actionable skill with real bundle files and clear reference navigation, weakened by verbosity and concept re-explanation across repetitive task sections. Tightening prose and moving detail into the existing reference files would lift conciseness and progressive disclosure.

Suggestions

Cut explanatory prose that restates what Claude already knows (e.g. what a screenshot or click is) and collapse the repetitive 'When to Use / How It Works / Best Practices' scaffolding shared across all seven tasks.

Replace the pseudocode agent-loop block with the real executable loop already present in scripts/form-filling-agent-template.py, or link to it directly.

Move the per-task action-dict API detail into references/computer-use-api.md so SKILL.md reads as a concise overview pointing one level deep.

Dimension	Reasoning	Score
Conciseness	Mostly efficient code examples are diluted by recurring prose that explains concepts Claude already knows ('Screenshots form the foundation of computer use', 'Computer use represents a major advancement') and repeated 'How It Works'/'Best Practices' boilerplate across seven near-identical task sections.	3 / 5
Actionability	Provides mostly executable action-dict and anthropic-client code that is close to copy-paste ready, but several blocks are explicitly labeled pseudocode ('Pseudo-code for agent loop') and the Quick Start leaves the agent loop as '...'.	4 / 5
Workflow Clarity	Multi-step workflows are clearly sequenced with screenshot-based verification checkpoints and an explicit error-recovery retry loop, though some validation is implicit rather than a hard gating step.	4 / 5
Progressive Disclosure	Good overall structure with three clearly signaled one-level-deep reference files (all real) plus a script template, but the ~565-line body inlines substantial task detail and API-style reference content that could live in the reference files.	4 / 5
	Total	15 / 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 well-constructed description that answers both 'what' and 'when' with concrete trigger phrases and a distinct computer-use niche. Slight room to sharpen actions beyond category labels and add common synonyms.

Dimension	Reasoning	Score
Specificity	Names the domain and several concrete capability areas ('UI automation, application control, and multi-step workflows', 'analyzing screen content'), but several are broad categories rather than atomic concrete actions, leaving minor coverage gaps.	4 / 5
Completeness	Clearly states what the skill does (advanced computer use patterns) and explicitly answers when to use it via a 'Use when...' clause listing concrete trigger scenarios.	5 / 5
Trigger Term Quality	Includes natural phrases users would say ('automating desktop tasks', 'testing applications', 'analyzing screen content', 'controlling software programmatically') with good coverage, but misses common synonyms/extensions like 'RPA' or 'screenshots'.	4 / 5
Distinctiveness Conflict Risk	Tied to a clear niche (Claude's computer use tool, Opus 4.5, zoom tool) with distinct triggers, though broad phrasing like 'automating desktop tasks' carries minor overlap risk with general automation skills.	4 / 5
	Total	17 / 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
skill_md_line_count	SKILL.md is long (570 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: fernandezbaptiste/Skrillz
Path: skills/claude-computer-use-advanced/SKILL.md
Commit: 93ed392

Reviewed: 1 day 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.