segment-anything-model

SAM: zero-shot image segmentation via points, boxes, masks.

Quality

64%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/mlops/models/segment-anything/SKILL.md

Quality

Content

72%

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

The content is highly actionable and well-structured with real reference files for progressive disclosure, but it is overlong with some redundancy and lacks validation checkpoints for its batch/destructive workflows.

Suggestions

Add explicit validation/verification checkpoints to batched inference and automatic mask generation workflows (e.g., check mask count/quality before proceeding, retry on empty output).

Tighten the document by consolidating the repeated point-prompt prediction examples that recur across the Quick start, Interactive segmentation, and Workflows sections.

Consider moving the ONNX deployment and some customized-generation detail into the advanced-usage reference to reduce main-file length while keeping the quick start lean.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient with executable code and minimal concept-padding, but its ~490-line length and repeated point-prompt prediction examples across several sections mean it could be tightened.	2 / 3
Actionability	It provides fully executable, copy-paste-ready code and commands with specific parameters (e.g., SamPredictor usage, ONNX export command, SamAutomaticMaskGenerator kwargs).	3 / 3
Workflow Clarity	Tasks are clearly sequenced by section, but batched inference and automatic mask generation lack explicit validation checkpoints or feedback loops, capping workflow clarity at 2 for batch/risky operations.	2 / 3
Progressive Disclosure	The overview points to two real, clearly signaled one-level-deep references (references/advanced-usage.md and references/troubleshooting.md), keeping detailed content appropriately split from the main file.	3 / 3
	Total	10 / 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 concrete and distinctive but terse: it states the capability and prompt modalities without any 'when to use' trigger guidance and with only partial coverage of natural user terms.

Suggestions

Add an explicit 'Use when...' clause listing triggers (e.g., 'Use when segmenting objects in images, generating masks, or building interactive annotation tools with point/box prompts').

Expand trigger terms to include natural phrasings users would say, such as 'segment images', 'mask generation', and 'extract objects'.

List a couple of concrete actions (e.g., 'generate object masks, refine with iterative prompts, auto-segment whole images') to lift specificity from domain+modalities to multiple distinct actions.

Dimension	Reasoning	Score
Specificity	The description names the domain ('zero-shot image segmentation') and concrete prompt modalities ('points, boxes, masks') but does not list multiple distinct actions, fitting the 'names domain and some actions' anchor.	2 / 3
Completeness	It clearly states what the skill does but includes no 'Use when...' trigger guidance, so the 'when' is missing and completeness is capped at 2 per the rubric.	2 / 3
Trigger Term Quality	'SAM', 'image segmentation', and 'zero-shot' are natural terms, but common user variations like 'segment images', 'mask generation', or 'extract objects' are missing, so coverage is partial.	2 / 3
Distinctiveness Conflict Risk	'SAM: zero-shot image segmentation via points, boxes, masks' carves a clear niche with distinctive triggers that are unlikely to conflict with other skills.	3 / 3
	Total	9 / 12 Passed

Validation

62%

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 — 10 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (507 lines); consider splitting into references/ and linking	Warning
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
relative_links	Relative link issues: 2 missing	Warning
referenced_paths_exist	Referenced path issues: 1 missing	Warning

	Total	10 / 16 Passed

Repository: NousResearch/hermes-agent
Path: skills/mlops/models/segment-anything/SKILL.md
Commit: d0d116b

Reviewed: 4 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.