CtrlK
BlogDocsLog inGet started
Tessl Logo

using-screenshots

Use this to capture screen context.

44

Quality

56%

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 ./.agency/plugins/nori/skills/using-screenshots/SKILL.md
SKILL.md
Quality
Evals
Security

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.

The body is a genuinely actionable, platform-thorough guide with copy-paste commands and a sensible error-recovery path. Its main weakness is token efficiency: the workflow is repeated three times and several sections explain things Claude already knows, roughly doubling the file's necessary length. A single canonical workflow section plus the error-handling and common-mistakes material would make it lean without losing coverage.

Suggestions

Collapse the 'Example Workflow' section and the Quick Reference 'Standard workflow' into the Step-by-Step Instructions, keeping one canonical pass instead of three renditions of the same commands.

Cut or compress content Claude already knows: the 'Analyze the Image' bullet list, the explanation of what uname -s returns, and flag-by-flag annotations like '-f specifies filename'.

Add a one-line verification step after capture (e.g., `ls -la /tmp/screenshot_*.png` or checking exit status) before invoking Read, since cancelled interactive selections can leave no file.

DimensionReasoningScore

Conciseness

The same workflow is stated three times (Quick Reference 'Standard workflow', Step-by-Step Instructions, and the Example Workflow section that re-lists every command), and sections like 'Analyze the Image' ('Identify UI elements', 'Spot visual bugs', 'Read text content') and 'Returns Darwin for macOS' explain things Claude already knows. This matches the score-2 anchor 'Noticeably verbose; several unnecessary explanations or padded sections'; it is above score 1 only because the bulk of the text is concrete commands rather than concept explanations.

2 / 5

Actionability

Every step has copy-paste-ready commands per platform (screencapture, gnome-screenshot -af, scrot -s, import), a tool-detection one-liner with fallback order, timestamped filenames, and even distro-specific install commands. This fully matches the score-5 anchor 'Fully executable; copy-paste ready code or commands; specific examples cover the common cases'.

5 / 5

Workflow Clarity

The six-step sequence is explicit and ordered, and there is an error-recovery branch ('Handling Missing Tools' with install suggestions and a manual fallback), matching 'Clear sequence with most checkpoints present'. It falls short of 5 because there is no checkpoint verifying the capture actually succeeded (e.g., confirming the file exists and is non-empty) before calling Read — and the interactive-selection flags can silently produce no file if the user cancels.

4 / 5

Progressive Disclosure

There are no bundle files and no references, but the skill is organized into clear, well-labeled sections (Overview, Quick Reference, Step-by-Step, error handling, Common Mistakes), which per the simple-skill guidance supports a high score. It is held to 4 rather than 5 because the Quick Reference / Step-by-Step / Example Workflow tripling is redundant structure — the same content presented three ways instead of one canonical pass.

4 / 5

Total

15

/

20

Passed

Description

25%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 a 7-word fragment that fails on nearly every axis: no concrete actions, no natural trigger terms, no 'when to use' guidance, and high overlap risk with anything screen-related. It reads like a placeholder rather than an engineered description, even though the body contains rich trigger phrases ('Take a screenshot', 'Look at my screen') that were never lifted into the frontmatter.

Suggestions

Add an explicit 'Use when...' clause listing natural triggers, e.g. 'Use when the user asks to take a screenshot, look at my screen, or analyze a UI bug visually.'

State concrete actions in third person, e.g. 'Captures screenshots via platform CLI tools (screencapture, gnome-screenshot, scrot, import) and loads them into context for visual analysis of UI elements and bugs.'

Include synonyms and concrete nouns (screenshot, screen capture, UI bug, .png) so the description both matches user phrasing and is distinct from generic image-handling skills.

DimensionReasoningScore

Specificity

"capture screen context" names the domain (screen capture) but offers only a single generic action with no specifics about platforms, tools, or analysis. This matches the score-2 anchor 'Names the domain but actions are minimal or generic' ("Processes PDF files") and falls short of score 3, which requires 1-2 concrete actions like 'takes screenshots and analyzes UI elements'.

2 / 5

Completeness

The 'what' is vague ("capture screen context") and there is no 'Use when...' clause or equivalent trigger guidance anywhere, matching the score-2 anchor 'Has a vague what and no when'. It cannot score 3 because the what is not clear and concrete, and the missing-when guideline independently caps it.

2 / 5

Trigger Term Quality

The only keyword is "screen context", which users rarely say; natural phrases like "take a screenshot", "look at my screen", "screenshot", or "UI bug" are absent. It sits at the score-2 anchor ('one or two generic keywords; missing the natural phrases users say') rather than score 3, which requires several relevant keywords.

2 / 5

Distinctiveness Conflict Risk

"Use this to capture screen context" is very broad and would plausibly trigger for any screen-, display-, UI-, or image-related request, matching the score-2 anchor 'Very broad; high overlap risk with many similar skills'. It is not score 3 because no concrete niche (screenshots via CLI, visual UI analysis) is articulated.

2 / 5

Total

8

/

20

Passed

Validation

100%

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
microsoft/FluidFramework
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.