scienceworld-target-locator

This skill determines the most likely location for a target object based on domain knowledge and environmental clues. Use when the agent needs to find a specific item (like an animal) but it is not in the current room. It analyzes the environment description and suggests a room to teleport to for further investigation.

Quality

78%

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

Fix and improve this skill with Tessl

tessl review fix ./experiments/src/skills/scienceworld/scienceworld-target-locator/SKILL.md

Quality

Content

77%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 tight, actionable single-purpose skill with a clear sequence, verification, and retry loop. Its key flaw is bundle hygiene: the existing reference file is unlinked and duplicated inline with conflicting and partially invalid room mappings, creating a correctness risk and wasted tokens.

Suggestions

Replace the inline heuristic table with a clearly signaled one-level reference to the bundle file, e.g. 'See [target_location_heuristics.md](references/target_location_heuristics.md) for the full object-to-room mappings and search priorities'.

Reconcile the body's mappings with the reference file — the body lists 'lab' for chemicals, which is not a valid room in the reference's room list, and the chemical mappings conflict outright.

Trim the example flow's redundant 'Context Check'/'Decision' lines or fold the accessibility check into the main procedure so it applies to every invocation, not just the example.

Dimension	Reasoning	Score
Conciseness	The body is lean (~28 lines) with a compact heuristic table and no explanation of concepts Claude already knows, but the example flow partially re-walks the procedure ('Context Check' and 'Decision' lines are largely trimmable), leaving minor instances of redundancy.	4 / 5
Actionability	Concrete commands are provided ('teleport to <ROOM>', 'look around') with a worked example ending in 'teleport to outside', but step 1 ('Classify the target object by type') is abstract and '<ROOM>' is a placeholder, so guidance is mostly rather than fully executable.	4 / 5
Workflow Clarity	The numbered procedure has an explicit verification checkpoint ('look around to verify the target is present') and a feedback loop ('If not, try the next likely room'), matching the clear-sequence-with-explicit-validation-and-retry anchor; no destructive or batch operations apply.	5 / 5
Progressive Disclosure	A bundle file (references/target_location_heuristics.md) exists with richer mappings, but the body never references it and instead inlines a smaller, partially conflicting table (chemicals map to 'lab, foundry' in the body vs 'bathroom, kitchen, workshop, greenhouse' in the reference, and 'lab' is not in the reference's room list) — content that should be separate is inline and the reference is buried.	3 / 5
	Total	16 / 20 Passed

Description

67%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 solid description with concrete third-person capability statements and an explicit 'Use when' trigger clause. Its main weakness is narrow trigger vocabulary and a somewhat generic 'find a specific item' framing that limits distinctiveness and keyword coverage.

Suggestions

Broaden trigger terms with natural variations users would say, e.g. 'locate an object', 'search for an item', or 'figure out where something is'.

Sharpen distinctiveness by naming the environment context (e.g. a game/simulation world with rooms) so it does not read as a generic object-search skill.

Dimension	Reasoning	Score
Specificity	Lists several concrete third-person actions — 'determines the most likely location for a target object', 'analyzes the environment description', 'suggests a room to teleport to' — with minor gaps in coverage, matching the 'several specific actions' anchor rather than the comprehensive 5.	4 / 5
Completeness	Both what ('determines the most likely location... analyzes the environment description and suggests a room to teleport to') and when ('Use when the agent needs to find a specific item... but it is not in the current room') are explicitly answered; the 'when' clause could be more specific with additional concrete trigger phrasings to reach 5.	4 / 5
Trigger Term Quality	Relevant keywords like 'find a specific item', 'not in the current room', and 'teleport' are present, but common variations users would say ('locate', 'search for', 'where is') are missing, fitting the 'some relevant keywords but missing synonyms' anchor.	3 / 5
Distinctiveness Conflict Risk	The teleport/room-navigation framing carves a fairly distinct niche with minimal overlap risk, but the generic opening 'find a specific item' could still collide with general search or navigation skills, placing it just below the clear-niche anchor.	4 / 5
	Total	15 / 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-target-locator/SKILL.md
Commit: 358ec82

Reviewed: 8 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.