CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

gpui-screenshots

Use whenever you need to see, click through or screenshot the GPUI client (crates/but-gpui), to check a UI change works, reproduce a UI bug, or show what a change looks like. Drives the app from a script in a window that never shows on the user's screen, against throwaway app data. Never start the app with `cargo run` or the plain binary to look at it.

Quality

95%

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

SKILL.md
Quality
Evals
Security

Quality

Content

96%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.

An exceptionally dense, expert-level body: every line is non-obvious project knowledge, guidance is copy-paste executable, and workflows carry explicit validation and feedback loops (especially the film-checking checklist). The only structural note is that with no bundle files, ~300 lines are inlined where the perf/power/film deep-dives could arguably live in one-level-deep references.

DimensionReasoningScore

Conciseness

Though ~300 lines, essentially every sentence carries non-inferable, project-specific knowledge (racy index entries after a copy, GPUI's stopped clock, AppKit napping after 30s, prompt fallback renderer ignoring Enter/Escape), and it assumes Claude's competence throughout with no padding or explanation of known concepts. This fits "every token earns its place" rather than the minor-trimming of the 4 anchor.

5 / 5

Actionability

Fully executable throughout: `crates/but-gpui/scripts/drive.sh run script.txt`, a complete mktemp/git-init throwaway-repo snippet, concrete script-line examples (add/wait/shot/click/key/type), `control.sh "select Branches" "pick first" "files on"`, a parameterized `film` example, and profiling commands like `sample <pid> <seconds> -file out.txt`. Copy-paste ready and covering the common cases, matching the 5 anchor.

5 / 5

Workflow Clarity

Multi-step flows are clearly sequenced (throwaway repo → build state → write script → run → read shots; step-at-a-time exploration; film workflow) with explicit validation checkpoints and feedback loops: "Read the shots to check them", "always `stop` it when done", "Read every FILM line. Don't grep for chosen frames", "After each fix, check everything that mustn't change", "compare means before calling a change a regression", "Say a motion is fixed only after filming the user's own steps". This matches the 5 anchor's validate-fix-retry pattern, not merely the mostly-checkpoints of 4.

5 / 5

Progressive Disclosure

No bundle files exist (no references/, scripts/, or assets/), so all ~300 lines sit inline under consistently clear, well-signaled section headers that make navigation easy. However, the specialized sections (Animations/film, Measuring performance, Measuring power) are plausible candidates for separate reference files and there is no overview pointing into them — good structure with minor organization gaps, the 4 anchor rather than the appropriately-split 5.

4 / 5

Total

19

/

20

Passed

Description

92%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 description: concrete third-person actions, an explicit "Use whenever..." trigger clause with natural phrases, and a tightly-scoped niche anchored to a specific crate path. The only weakness is trigger-term breadth — a user saying "screenshot the GitButler app" or "preview the UI" would not literally match the wording.

DimensionReasoningScore

Specificity

Lists multiple concrete actions in third person — "see, click through or screenshot the GPUI client (crates/but-gpui)", "check a UI change works, reproduce a UI bug, or show what a change looks like", "Drives the app from a script" — with comprehensive coverage of the skill's use cases; no gaps to point to, so it fits the top anchor rather than 4.

5 / 5

Completeness

Explicitly answers both questions: "Use whenever you need to see, click through or screenshot... to check a UI change works, reproduce a UI bug, or show what a change looks like" gives concrete trigger phrases, and the what is stated concretely (drives the app from a script in an off-screen window against throwaway app data, plus the prohibition on `cargo run`). The 'when' is explicit with concrete triggers, so above the 4 anchor.

5 / 5

Trigger Term Quality

Natural phrases users would say are present ("screenshot", "see", "click through", "UI bug", "UI change"), but coverage stays tied to "GPUI client (crates/but-gpui)" and misses common user-worded synonyms such as "GitButler", "the app", or "preview" — good coverage with a few natural terms missing, not the comprehensive synonym coverage of the 5 anchor.

4 / 5

Distinctiveness Conflict Risk

Names an exact crate path (crates/but-gpui) and a distinct niche — scripted, off-screen driving of one specific client with throwaway data — so it would not plausibly trigger for any other skill. Clear niche with distinct triggers, matching the 5 anchor.

5 / 5

Total

19

/

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

gitbutlerapp/gitbutler

Path

.agents/skills/gpui-screenshots/SKILL.md

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.