CtrlK
BlogDocsLog inGet started
Tessl Logo

scienceworld-target-identifier

Analyzes room observations to identify objects matching a given target description (e.g., 'living thing'). Triggered after exploring a room when the agent needs to locate a specific type of item. Processes the observation list, filters objects based on target criteria, and returns candidate objects for further action.

62

Quality

78%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./experiments/src/skills/scienceworld/scienceworld-target-identifier/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

63%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.

The skill body is well-organized, concise, and sequence-clear with useful error-recovery guidance, but its actionability is undermined by a dangling reference to a nonexistent classification script and vague semantic-matching direction. Bundle hygiene is the weakest area: one referenced resource is missing, another is an unreferenced truncated stub, and the taxonomy is not linked by path.

Suggestions

Resolve the dangling "bundled classification script" reference in Step 2 — either add the actual script under scripts/ or inline the concrete classification rules (e.g., the living-thing criteria from references/taxonomy.md) so the filtering step is executable as written.

Reference bundle files explicitly by path (e.g., "See [references/taxonomy.md](references/taxonomy.md)") instead of vague phrases like "bundled resources", and complete or remove the truncated references/action_patterns.md stub.

Replace "Use semantic similarity matching" with a concrete procedure (e.g., specific heuristics or the exact taxonomy categories to match against) so generic target descriptions are handled executably.

DimensionReasoningScore

Conciseness

The body is efficient and domain-specific (observation parsing cues, prioritization rules, environment rules like "All containers are open"), with no explanation of concepts Claude already knows. Minor trimmable padding exists — the Purpose section restates the frontmatter description, and phrases like "for your current task" add little — keeping it just below anchor 5.

4 / 5

Actionability

There is genuinely concrete guidance (exact observation markers "Here you see:", "a substance called", exact-name rules, and a worked example ending in `focus on turtle egg` → `pick up turtle egg` → `teleport to bathroom`), but the core Step 2 defers to "the bundled classification script", which does not exist in the bundle (no scripts/ directory), and "Use semantic similarity matching" is unspecified. A missing central executable resource is a key-detail gap, matching anchor 3 rather than anchor 4's "minor gaps".

3 / 5

Workflow Clarity

Steps 1–4 are clearly sequenced, and the Error Handling section provides recovery loops ("If no matches found: Teleport to another room and repeat", "If ambiguous matches: Use `examine OBJ`", "If classification uncertain: Check reference taxonomy"), approaching anchor 5. It stays at 4 because checkpoints are a separate section rather than embedded validation steps in the workflow, and the fallback path relies on a taxonomy reference and script that are poorly wired up.

4 / 5

Progressive Disclosure

The body is well-sectioned, but reference signaling is weak: the taxonomy is referenced only vaguely as "reference taxonomy in bundled resources" with no path or link, the "bundled classification script" points to a nonexistent file, and references/action_patterns.md is a truncated stub (ends mid-outline at "### Pattern 1: Simple Pickup") that is never referenced at all. This matches anchor 3 — references present but not clearly signaled — rather than anchor 4's "references mostly clear".

3 / 5

Total

14

/

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 strong description that clearly states both capability and trigger conditions with mostly concrete language. Its main weaknesses are slightly generic filtering language and the absence of the environment name (ScienceWorld), which would sharpen distinctiveness.

DimensionReasoningScore

Specificity

The description lists several concrete actions — "Analyzes room observations to identify objects", "Processes the observation list", "filters objects based on target criteria", and "returns candidate objects" — anchored to a clear domain with an example target ("living thing"). It falls short of anchor 5 because "filters objects based on target criteria" is generic and coverage of the filtering/prioritization behavior is incomplete.

4 / 5

Completeness

It explicitly answers both questions: the "what" is the analyze/process/filter/return pipeline, and the "when" is a concrete trigger clause — "Triggered after exploring a room when the agent needs to locate a specific type of item". This matches anchor 5's requirement for explicit what-and-when with concrete trigger phrases; anchor 4 would require the when to be less explicit than it is.

5 / 5

Trigger Term Quality

Natural trigger phrases are present: "Triggered after exploring a room", "locate a specific type of item", "living thing", and "room observations" — terms a user or task would plausibly use. It misses common variations like "find", "search", or "look around", keeping it below anchor 5.

4 / 5

Distinctiveness Conflict Risk

The niche is fairly clear (room-observation object identification with target-description filtering), but the description never names the ScienceWorld environment, so "room observations" and "locate objects" could overlap with other environment-navigation or search skills — a minor overlap risk consistent with anchor 4 rather than anchor 5's "clear niche with minimal conflict risk".

4 / 5

Total

17

/

20

Passed

Validation

100%

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
Reviewed

Table of Contents

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.