segmentation-sam2

Interactive click-to-segment using Segment Anything 2 — AI-assisted labeling for Annotation Studio

Quality

71%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./skills/segmentation/sam2-segmentation/SKILL.md

Quality

Content

85%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is well-structured and highly actionable, with concrete bidirectional protocol examples and a clear install path. Its only real weakness is mild marketing redundancy in the intro that slightly undercuts token efficiency.

Suggestions

Tighten the intro and 'What You Get' bullets by dropping marketing language ('pixel-perfect', 'full integration') and avoiding repetition of the segmentation capability.

Add a brief note on protocol error/feedback handling (e.g. an error event shape) so Claude knows how to recover from a failed segmentation request.

Clarify that deploy.sh is the bootstrapper and confirm it is bundled, since the body references ./deploy.sh but no such file is present in the bundle.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient with concrete protocol examples and no basic-concept padding, but the intro and 'What You Get' bullets carry mild marketing fluff ('pixel-perfect masks', 'full integration') and overlap with each other, so it could be tightened to reach the top anchor.	2 / 3
Actionability	Concrete, copy-paste-ready JSON lines examples are shown for both stdin and stdout directions plus an exact install command, giving fully executable guidance rather than vague direction.	3 / 3
Workflow Clarity	For this single-purpose interactive skill the request/response sequence (frame → segment → track → stop) is laid out concretely and unambiguously, satisfying the simple-skill carve-out for a top score without needing destructive-operation checkpoints.	3 / 3
Progressive Disclosure	At under 40 lines with well-organized sections (What You Get, Protocol, Installation) and a real referenced bundle path (scripts/segment.py), the structure earns the top anchor for a simple skill with no need for nested reference docs.	3 / 3
	Total	11 / 12 Passed

Description

57%

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 specific and clearly niched to SAM2 within Annotation Studio, but it omits explicit 'when to use' trigger guidance and only partially covers natural trigger-term variations. Adding a 'Use when...' clause would raise the weakest dimensions.

Suggestions

Add an explicit trigger clause, e.g. 'Use when the user wants to segment objects in frames, create masks, or speed up annotation labeling in Annotation Studio.'

Expand trigger terms with common user phrasings such as 'segment objects', 'generate masks', 'click-to-segment video frames', and 'SAM2'.

Optionally enumerate a second concrete action (e.g. mask export or tracking) to push specificity toward the top anchor.

Dimension	Reasoning	Score
Specificity	Names the core action ('Interactive click-to-segment') and domain ('AI-assisted labeling for Annotation Studio'), but it is essentially one capability rather than the multiple distinct concrete actions the top anchor lists.	2 / 3
Completeness	It clearly states what the skill does but contains no 'Use when...' clause or equivalent trigger guidance, so per the rubric it caps at 2 rather than 3.	2 / 3
Trigger Term Quality	Relevant natural keywords are present ('click-to-segment', 'Segment Anything 2', 'labeling', 'Annotation Studio'), but common user variations like 'segment', 'mask', or 'click to segment objects' are only partly covered.	2 / 3
Distinctiveness Conflict Risk	The reference to a specific model ('Segment Anything 2') and a specific platform ('Annotation Studio') gives it a clear niche that is unlikely to trigger for unrelated skills.	3 / 3
	Total	9 / 12 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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: SharpAI/DeepCamera
Path: skills/segmentation/sam2-segmentation/SKILL.md
Commit: 2264fcb

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.