scienceworld-object-focuser

This skill selects and focuses on a specific object to signal task intent or prepare it for manipulation. Use when you have identified a target object that meets task criteria (e.g., a living thing) and need to formally select it before performing actions like moving or using the object in ScienceWorld tasks. The skill uses the 'focus on OBJ' action, taking the object name as input.

Quality

81%

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

Quality

Content

78%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 concise, actionable, well-sequenced body for a simple single-action skill, with copy-paste-ready commands and a confirmation checkpoint. The main weakness is progressive disclosure: a useful reference file exists but is not linked from the body, leaving richer patterns and anti-patterns undiscoverable.

Suggestions

Add a brief "See also" or "More examples" line linking to references/usage_examples.md so the additional patterns (re-focusing, implicit focus, anti-patterns like wrong object names) are discoverable from the body.

Add a short feedback-loop note for the failure case where focus fails or the object name is wrong (e.g., re-examine and retry with the exact observation noun phrase) to raise workflow clarity.

Trim the redundancy between the Core Instruction steps and the worked Example, or reframe the Example as a distinct second scenario to avoid restating the same three steps.

Dimension	Reasoning	Score
Conciseness	The body is lean and avoids explaining concepts Claude already knows, with each section earning its place; held at 4 rather than 5 because the worked Example restates the Core Instruction steps and "When to Use" overlaps with the description, giving minor trimmable redundancy.	4 / 5
Actionability	Provides the exact executable command form `focus on <OBJECT_NAME>` with concrete object-name examples (`dove egg`, `copper wire`, `beaker`) and a full worked action sequence (`look around` -> `focus on dove egg` -> `pick up dove egg`), which is copy-paste ready within the ScienceWorld action space and covers common cases.	5 / 5
Workflow Clarity	Core Instruction gives a clear 3-step sequence (Identify Target -> Execute Focus -> Proceed) with a confirmation checkpoint ("After receiving a confirmation observation, continue"); not 5 because there is no explicit error-recovery feedback loop for failure modes like an incorrect object name or a failed focus, which the reference file actually documents.	4 / 5
Progressive Disclosure	The body is well-organized into clear sections (Purpose, When to Use, Core Instruction, Key Considerations, Example), but the existing bundle file references/usage_examples.md — which contains valuable extra examples, common patterns, and anti-patterns — is never linked or signaled from the body, so the reference is present but not clearly surfaced.	3 / 5
	Total	16 / 20 Passed

Description

83%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 well-constructed description that clearly states both what the skill does and when to use it, with concrete action syntax and domain-appropriate trigger terms. Its main limitation is that it covers a single simple action, which caps specificity and trigger breadth.

Dimension	Reasoning	Score
Specificity	Names the concrete action ("focus on OBJ"), its input (object name), and its purpose (signal task intent / prepare for manipulation) with examples like "moving or using the object"; falls short of 5 because it is a single action rather than a comprehensive set of actions, and above 3 because the action and mechanism are specified concretely.	4 / 5
Completeness	Explicitly answers both "what" ("selects and focuses on a specific object... uses the 'focus on OBJ' action, taking the object name as input") and "when" ("Use when you have identified a target object that meets task criteria... before performing actions like moving or using the object") with concrete trigger phrases.	5 / 5
Trigger Term Quality	Includes natural domain terms a user/agent would say ("focus on", "target object", "ScienceWorld tasks", "living thing", "moving or using the object"); not 5 because it lacks synonym/extension coverage, and above 3 because the actual action verb and phrasing are present rather than only generic keywords.	4 / 5
Distinctiveness Conflict Risk	Scoped to the specific "focus on OBJ" action within ScienceWorld, giving it a clear niche with minimal conflict risk against unrelated skills; held at 4 rather than 5 because it sits within a family of related ScienceWorld manipulation skills (pick up, move) where some overlap is possible.	4 / 5
	Total	17 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: zjunlp/SkillNet
Path: experiments/src/skills/scienceworld/scienceworld-object-focuser/SKILL.md
Commit: 076c6df

Reviewed: 6 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.