peekaboo

Capture and automate macOS UI with the Peekaboo CLI.

1.92x

Quality

66%

Does it follow best practices?

Impact

77%

1.92x

Average score across 3 eval scenarios

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./openclaw/skills/peekaboo/SKILL.md

The canonical home for this skill is peekaboo in Hung-Reo/hungreo-openclaw

Quality

Content

80%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 an exceptionally lean and actionable single-file catalog: every command is a one-liner and the examples are copy-paste ready. Its weaknesses are the absence of validation/error-recovery checkpoints in the recommended flows and the fact that the entire command and parameter reference is inlined in SKILL.md rather than split into reference files.

Suggestions

Add validation checkpoints and error-recovery guidance to the recommended flow, e.g. 'If `click --on B3` fails, re-run `see --annotate` to refresh element IDs and confirm the element exists' and a permissions pre-flight check before automating.

Move the full command catalog and common-parameter tables into a references/ file (e.g. COMMANDS.md), keeping SKILL.md to the overview, quickstart, and the most common flows with a clearly signaled 'See COMMANDS.md' link.

State how to verify results after acting (e.g. re-capture or query the element state after click/type) so the automation loop has an explicit feedback step.

Dimension	Reasoning	Score
Conciseness	Every section is a terse one-line-per-command catalog with copy-paste examples and no explanation of concepts Claude already knows — matching 'Lean and efficient; assumes Claude's competence; every token earns its place'. It is not score 4 because there is no padding to trim: even the parameter sections are compact flag lists rather than prose.	5 / 5
Actionability	The body provides fully executable, copy-paste-ready commands throughout ('peekaboo see --app Safari --window-title "Login" --annotate --path /tmp/see.png', the Quickstart happy path, and eight worked example blocks), fitting 'Fully executable; copy-paste ready code or commands; specific examples cover the common cases'. It is not score 4 because the examples cover the realistic common cases with real flag combinations rather than leaving minor gaps.	5 / 5
Workflow Clarity	Sequences are present and the recommended flow is labeled ('See -> click -> type (most reliable flow)' and the note 'Use `peekaboo see --annotate` to identify targets before clicking'), but validation checkpoints and error-recovery guidance are missing — e.g. what to do when `see` finds no matching element, when permissions are missing (only 'Requires Screen Recording + Accessibility permissions' is stated), or how to verify a click landed. This matches 'Steps listed but validation gaps; sequence present but checkpoints missing or implicit' rather than score 4, which requires most checkpoints present.	3 / 5
Progressive Disclosure	No bundle files (references/, scripts/, assets/) exist, and the body inlines the entire ~60-command catalog plus common-parameter tables in one file with good headers, but no one-level-deep references to offload detail — matching 'Some structure but could be better organized; content that should be separate is inline'. It is above score 2 because the file is well-sectioned and navigable (not a header-less wall of text), and below score 4 because the command catalog and parameter tables are exactly the bulk that belongs in a separate reference file for a CLI this large.	3 / 5
	Total	16 / 20 Passed

Description

53%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 concise, third-person, and clearly scoped to macOS UI automation with a named tool, but it stops at a high-level what. It lacks a 'Use when...' trigger clause and natural trigger vocabulary (screenshots, click, type, window management), which limits completeness and trigger-term quality.

Suggestions

Add an explicit trigger clause, e.g. 'Use when the user asks to capture the screen, click or type into macOS apps, or automate any native Mac UI.'

Enumerate the concrete capability groups (capture screenshots, drive mouse/keyboard input, manage apps/windows/menus) so the 'what' is comprehensive rather than two generic verbs.

Include natural trigger terms users would actually say — 'screenshot', 'screen capture', 'UI automation', 'click a button', 'AppleScript alternative' — to improve trigger-term coverage.

Dimension	Reasoning	Score
Specificity	The description names the domain ("macOS UI") and two actions ("Capture and automate") with the concrete tool name ("the Peekaboo CLI"), matching the anchor 'Names domain and 1-2 concrete actions, but not comprehensive'. It is not score 4 because 'automate' is generic and there is no enumeration of concrete actions (click, type, screenshots, window management); not score 2 because 'capture' plus the named CLI goes beyond a minimal, action-free statement.	3 / 5
Completeness	There is a clear 'what' ("Capture and automate macOS UI with the Peekaboo CLI") but no 'Use when...' clause or equivalent trigger guidance, so per the judging guidelines completeness is capped at 3 ('Has a clear what but when is missing or only weakly implied'). It is not score 4 because the 'when' is entirely absent rather than merely under-specified.	3 / 5
Trigger Term Quality	Relevant keywords like "Capture", "automate", "macOS UI", and "Peekaboo CLI" are present, but common natural phrases users would say — "screenshot", "screen capture", "click", "UI automation", "AppleScript alternative" — are missing, fitting the anchor 'Some relevant keywords but missing common variations or synonyms'. It is above score 2 because more than one or two relevant, domain-specific keywords appear; below score 4 because the natural vocabulary users actually use when asking for this is largely absent.	3 / 5
Distinctiveness Conflict Risk	"Capture and automate macOS UI with the Peekaboo CLI" carves out a distinct niche (macOS-only UI automation via a named CLI), fitting 'Mostly distinct; minor overlap risk with closely related skills'. It is below score 5 because the anchor-5 example includes explicit distinct trigger phrases, which this description lacks; above score 3 because it is far more specific than generic 'Works with document files' phrasing and unlikely to fire for the wrong skill.	4 / 5
	Total	13 / 20 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: trpc-group/trpc-agent-go
Path: openclaw/skills/peekaboo/SKILL.md
Commit: 9bb1c42

Reviewed: 5 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.