CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

segmentation-sam2

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

Quality

68%

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 ./skills/segmentation/sam2-segmentation/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

68%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 concise and actionably concrete with solid protocol examples and clean sectioning, but it lacks explicit validation/recovery steps for a batch video-tracking workflow and references a deploy.sh that is absent from the bundle. Adding a validation checkpoint and verifying the deploy reference would close the main gaps.

Suggestions

Add an explicit validation/recovery step for the tracking loop, e.g. check that each 'tracked' response has a mask_path and retry or report on missing frames before proceeding.

Resolve the missing bundle reference: either include deploy.sh in the bundle or correct the Installation section to point at an existing bootstrapper.

Clarify the error/stop protocol (e.g. what the skill emits on failure to segment) so the sequence has a complete feedback loop.

DimensionReasoningScore

Conciseness

The body is lean and well-paced: a one-line intro, a tight feature list, concrete JSONL protocol examples, and a single install command, with only minor phrasing that could be trimmed; it avoids explaining concepts Claude already knows.

4 / 5

Actionability

Provides concrete, copy-paste-ready JSONL examples for both directions of the protocol (frame/segment/track/stop and ready/encoded/segmented/tracked) plus the `./deploy.sh` command, covering the common cases; held at 4 rather than 5 only because it is a protocol spec rather than directly runnable code with full edge-case handling.

4 / 5

Workflow Clarity

The message sequence is implied by the protocol examples (ready -> frame -> segment -> tracked) but there are no explicit validation checkpoints or error-recovery feedback loops; since video tracking is a batch operation, the missing validation cap holds this at 3.

3 / 5

Progressive Disclosure

Well-organized into clear sections (What You Get, Protocol, Installation) with the implementation correctly pushed to scripts/segment.py and a clearly signaled install reference; the referenced deploy.sh is missing from the bundle, a minor organization gap that keeps it just below 5.

4 / 5

Total

15

/

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 clearly conveys what the skill does and occupies a distinct niche, but it lacks an explicit 'Use when...' trigger clause and only lightly covers natural user phrasing. Adding trigger guidance and a few synonyms would lift completeness and trigger-term quality.

Suggestions

Append an explicit trigger clause, e.g. 'Use when the user wants to segment or mask objects in video frames, click-to-segment for labeling, or mentions SAM2 / Segment Anything 2.'

Add common natural synonyms users say, such as 'mask', 'pixel masks', and 'SAM2', to broaden trigger-term coverage.

Optionally name one or two more concrete actions (e.g. 'propagate masks across video frames') to move specificity toward 4-5.

DimensionReasoningScore

Specificity

Names the domain ('segmentation') and one concrete action ('Interactive click-to-segment') plus the purpose ('AI-assisted labeling'), but does not enumerate multiple specific actions; fits the '1-2 concrete actions, not comprehensive' anchor.

3 / 5

Completeness

Provides a clear 'what' (click-to-segment with SAM2 for labeling) but has no 'Use when...' or equivalent explicit trigger guidance; per the guidelines a missing trigger clause caps completeness at 3.

3 / 5

Trigger Term Quality

Includes relevant natural terms ('segment', 'labeling', 'annotation') and the product name 'Segment Anything 2', but omits common synonyms/extensions users might say ('mask', 'SAM2', 'pixel-perfect masks' as a trigger); matches 'some relevant keywords but missing common variations'.

3 / 5

Distinctiveness Conflict Risk

Targets a clear niche (SAM2 model + Annotation Studio integration) unlikely to collide with unrelated skills, but lacks explicit distinct trigger phrases that would push it to 5; sits above the 'somewhat specific' anchor at 3.

4 / 5

Total

13

/

20

Passed

Validation

93%

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

CriteriaDescriptionResult

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

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.