egohos-segmentation

Egocentric Hand-Object Segmentation (EgoHOS) - pixel-level hand and object segmentation in egocentric videos. Outputs fine-grained segmentation masks with hand regions highlighted. Specialized for hand-object interaction scenarios with pixel-accurate masks. Ideal for detailed interaction analysis.

Quality

73%

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/vision/egohos-segmentation/SKILL.md

Quality

Content

65%

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 with concrete executable code, but it is verbose and presents everything inline with no progressive disclosure or validation feedback loops for its batch/destructive operations. Splitting detail into reference files, deduplicating the install steps, and adding verification checkpoints would lift the weaker dimensions.

Suggestions

Add explicit validation/verification steps for batch and video-processing workflows (e.g., confirm masks were written, check output video integrity, handle empty-hand/no-object frames) with a validate→fix→retry loop, which would lift workflow_clarity.

Move detailed material (model specifications, the four full usage examples, integration/limitations) into one-level-deep reference files (e.g., EXAMPLES.md, MODEL_SPECS.md) and keep SKILL.md as a concise overview, improving progressive_disclosure and conciseness.

Deduplicate the repository-clone and dependency-install instructions that appear in both Section 2 and the Installation section, and reconcile the inconsistent model-download and checkpoint filenames.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient with concrete code, but it is verbose: the git-clone/install sequence is duplicated (Section 2 and the Installation section) and prose such as the Overview and 'Key advantage' explains concepts Claude already knows. It could be tightened considerably without losing clarity.	2 / 3
Actionability	Provides fully executable bash commands and multiple copy-paste-ready Python examples (demo.py invocation, mask format, batch processing, ROI extraction), matching the top anchor for concrete executable guidance. Minor path/filename inconsistencies (download_models.sh vs download_pretrained_models.py, best_model.pth vs model.pth) are present but do not make the code non-executable.	3 / 3
Workflow Clarity	Steps are listed for setup and usage, but there are no explicit validation checkpoints or validate→fix→retry feedback loops, and batch/destructive operations (video and dataset batch processing) lack verification. The rubric caps workflow_clarity at 2 when validation is missing in these contexts.	2 / 3
Progressive Disclosure	The skill is a monolithic single file with no bundle references and all content inline; section organization exists, but reference material (model specs, multiple full examples, integration notes) that could live one level deep is kept inline. This matches the 'some structure but content that should be separate is inline' anchor rather than the top anchor.	2 / 3
	Total	9 / 12 Passed

Description

82%

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 trigger-rich with a clear niche, but it lacks an explicit 'Use when...' clause, so the when-to-use guidance remains implicit rather than stated. Adding an explicit trigger sentence would raise completeness to the top level.

Suggestions

Add an explicit 'Use when ...' sentence naming the natural trigger phrases users would say (e.g., 'Use when segmenting hands and objects in first-person/egocentric videos, generating hand-object masks, or analyzing contact regions').

Tighten redundant capability phrasing ('pixel-accurate masks' repeats the segmentation idea) to keep the description concise.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions such as 'pixel-level hand and object segmentation in egocentric videos', 'Outputs fine-grained segmentation masks with hand regions highlighted', and 'pixel-accurate masks', matching the top anchor for multiple specific concrete actions.	3 / 3
Completeness	Clearly states what the skill does, but 'when' guidance is only implied via phrases like 'Specialized for' and 'Ideal for'; there is no explicit 'Use when...' clause, which the guidelines say caps completeness at 2.	2 / 3
Trigger Term Quality	Uses natural terms a user would say — 'Egocentric Hand-Object Segmentation', 'pixel-level', 'segmentation masks', 'egocentric videos', and 'hand-object interaction' — giving good coverage of natural language triggers.	3 / 3
Distinctiveness Conflict Risk	Occupies a clear niche (egocentric hand-object segmentation with pixel-accurate masks) with distinct triggers, making it unlikely to fire for the wrong skill.	3 / 3
	Total	11 / 12 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
referenced_paths_exist	Referenced path issues: 2 missing	Warning

	Total	13 / 16 Passed

Repository: wu-yc/LabClaw
Path: skills/vision/egohos-segmentation/SKILL.md
Commit: df37802

Reviewed: about 1 hour 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.