segment-anything-model

Foundation model for image segmentation with zero-shot transfer. Use when you need to segment any object in images using points, boxes, or masks as prompts, or automatically generate all object masks in an image.

Quality

78%

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 ./backend/cli/skills/llm-tools/segment-anything/SKILL.md

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.

A highly actionable, code-rich skill body with clear structure and well-signaled references, weakened by verbosity and duplication between SKILL.md and the reference files and by workflows lacking validation checkpoints for batch operations.

Suggestions

Deduplicate SKILL.md against references/advanced-usage.md (batched inference, performance optimization, ONNX) so the main file is an overview that points to details — this also improves conciseness.

Add explicit validation/verification steps to batched and destructive workflows (e.g., assert mask shape/coverage, verify predicted_iou threshold before accepting batch results) to lift workflow_clarity above 3.

Remove the redundant mask-data-structure block (keep it in one place) and trim the ASCII architecture diagram, which explains a concept Claude already knows.

Dimension	Reasoning	Score
Conciseness	Mostly efficient with concrete code, but the ~490-line body repeats content (mask data structure appears in both 'Automatic mask generation' and 'Output format'; batched inference and performance optimization are duplicated in references/advanced-usage.md) and includes an ASCII architecture diagram explaining a concept Claude already knows, fitting the 'mostly efficient but could be tightened' anchor.	3 / 5
Actionability	Fully executable, copy-paste-ready code throughout — install commands, checkpoint wget URLs, complete SamPredictor/HuggingFace/automatic-mask/ONNX examples covering the common cases, matching the top anchor.	5 / 5
Workflow Clarity	Workflows are presented as single-purpose code recipes rather than explicitly sequenced multi-step processes, and batched inference (a batch operation) lacks any validation/verification checkpoint, so per the feedback-loop cap it cannot exceed 3 even though quick-start sequencing is reasonable.	3 / 5
Progressive Disclosure	Clear section structure with two well-signaled, one-level-deep references (advanced-usage.md, troubleshooting.md) linked from a References section, but significant content is inlined/duplicated in SKILL.md that already lives in the bundle files, leaving minor organization gaps below the top anchor.	4 / 5
	Total	15 / 20 Passed

Description

87%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, third-person description that clearly states what SAM does and when to use it with concrete trigger phrases and a distinct niche. Minor improvement would come from adding synonyms and file extensions to broaden natural trigger coverage.

Suggestions

Add common synonyms and concrete terms users say (e.g., 'instance segmentation', 'mask generation', '.jpg/.png images') to lift trigger_term_quality from good to comprehensive.

Mention an output-oriented action verb (e.g., 'extract object masks' or 'cut out objects') to round out specificity coverage.

Dimension	Reasoning	Score
Specificity	Names the domain and several concrete actions — 'segment any object in images using points, boxes, or masks as prompts' and 'automatically generate all object masks' — with only minor gaps in coverage (no output formats or extraction verbs), fitting the 'several specific actions' anchor below the comprehensive 5.	4 / 5
Completeness	Explicitly answers both 'what' ('Foundation model for image segmentation with zero-shot transfer') and 'when' ('Use when you need to segment any object... or automatically generate all object masks') with concrete trigger phrases, matching the top anchor.	5 / 5
Trigger Term Quality	Good keyword coverage with natural phrases users would say ('image segmentation', 'segment any object', 'object masks', 'points, boxes, or masks'), but missing common synonyms and file extensions, so it sits below the comprehensive 5 anchor.	4 / 5
Distinctiveness Conflict Risk	Clear niche (zero-shot prompted image segmentation) with distinct triggers and minimal overlap risk with detection/categorization tools, matching the 'clear niche with distinct triggers' anchor.	5 / 5
	Total	18 / 20 Passed

Validation

75%

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 — 12 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (502 lines); consider splitting into references/ and linking	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	12 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/llm-tools/segment-anything/SKILL.md
Commit: fcd2e11

Reviewed: 2 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.