CtrlK
BlogDocsLog inGet started
Tessl Logo

computer-use-agents

Build AI agents that interact with computers like humans do - viewing screens, moving cursors, clicking buttons, and typing text. Covers Anthropic's Computer Use, OpenAI's Operator/CUA, and open-source alternatives. Critical focus on sandboxing, security, and handling the unique challenges of vision-based control. Use when: computer use, desktop automation agent, screen control AI, vision-based agent, GUI automation.

51

Quality

57%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./.agent/skills/computer-use-agents/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

26%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.

The SKILL.md body is an empty shell: an empty '## Patterns' section and three links to sub-skill files that are absent from the bundle. While token-efficient on its face, it provides zero actionable guidance, no workflow, and broken navigation, making the skill unusable as written.

Suggestions

Add substantive, executable content — e.g., a minimal Anthropic Computer Use tool-definition and tool-use loop code example, and concrete sandboxing guidance — either inline or in the referenced sub-skill files.

Create the three referenced files under ./sub-skills/ (perception-reasoning-action-loop.md, sandboxed-environment-pattern.md, anthropic-computer-use-implementation.md) or remove the broken links, since none of them exist in the current bundle.

Remove the empty '## Patterns' section and replace it with a brief overview that tells the reader when to consult each knowledge module.

DimensionReasoningScore

Conciseness

The body is extremely lean with no padding or explanation of concepts Claude already knows, so it is far above the verbose anchors; however, the empty "## Patterns" heading contributes nothing and could be trimmed, keeping it just below the every-token-earns-its-place anchor 5. It is not a 3 because there is no unnecessary explanation at all — only a wasted heading.

4 / 5

Actionability

The body contains no concrete code, commands, or guidance of any kind — only a title and three links to sub-skill files (which do not exist in the bundle). It entirely fails to instruct, matching the anchor 'entirely vague or abstract; no concrete code or commands'. It cannot be a 2 because there are not even high-level hints or steps toward executing the skill.

1 / 5

Workflow Clarity

No steps or sequence are present anywhere in the body; the '## Patterns' section is empty and the numbered items are bare links to nonexistent files. This matches anchor 1 ('steps missing or incoherent; no sequence') and is clearly below the rough-sequence-with-gaps anchor 2.

1 / 5

Progressive Disclosure

The reference structure is nominally one level deep and clearly signaled, but the three referenced ./sub-skills/*.md files do not exist in the bundle, so navigation fails entirely and all substantive content is missing. This lands at anchor 2 (minimal structure, references effectively unusable) rather than 3, because there is no overview content at all to organize — the body is an empty shell plus broken links.

2 / 5

Total

8

/

20

Passed

Description

88%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.

The description is strong: it states concrete capabilities, covers specific vendor APIs, and includes an explicit 'Use when' clause with natural trigger phrases. The only weaknesses are slightly unnatural trigger phrasing ("screen control AI") and a few broad terms that create minor overlap risk with generic automation skills.

DimensionReasoningScore

Specificity

The description lists multiple concrete actions ("viewing screens, moving cursors, clicking buttons, and typing text") plus concrete focus areas ("sandboxing, security", "vision-based control") and named systems (Anthropic's Computer Use, OpenAI's Operator/CUA), giving comprehensive coverage of the domain. It is not merely one level down (4) because there are no significant gaps in the action coverage for the computer-use agent domain.

5 / 5

Completeness

It explicitly answers both questions: what ("Build AI agents that interact with computers like humans do - viewing screens, moving cursors, clicking buttons, and typing text") and when ("Use when: computer use, desktop automation agent, screen control AI, vision-based agent, GUI automation") with concrete trigger phrases. This matches the anchor-5 example structure of explicit what + explicit when with triggers, not the anchor-4 case where 'when' could be more specific.

5 / 5

Trigger Term Quality

"Use when: computer use, desktop automation agent, screen control AI, vision-based agent, GUI automation" provides good natural keyword coverage, but phrases like "screen control AI" are unnatural and common variations such as "browser automation" or "computer use agent" are missing. It falls between the good-coverage (4) and comprehensive-with-synonyms (5) anchors, closer to 4.

4 / 5

Distinctiveness Conflict Risk

The niche is clear (vision-based computer-use agents, with named APIs like Anthropic Computer Use and OpenAI Operator/CUA), but broader trigger terms such as "GUI automation" and "desktop automation agent" carry minor overlap risk with generic automation or testing skills. Mostly distinct with minor overlap risk fits anchor 4 better than the minimal-conflict anchor 5.

4 / 5

Total

18

/

20

Passed

Validation

87%

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

CriteriaDescriptionResult

frontmatter_unknown_keys

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

Warning

relative_links

Relative link issues: 3 missing

Warning

Total

14

/

16

Passed

Repository
Dokhacgiakhoa/Agent-Skills-4-Vibe-Coding-CLI
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.