prioritize-reference-files-6e4111

Always read and use provided reference files for data before attempting external searches or fabricating information

Quality

62%

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 ./benchmarks/gdpval/skills/prioritize-reference-files-6e4111/SKILL.md

Quality

Content

67%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 body is well-structured with concrete, mostly-executable examples and a clear step sequence plus a checklist. Its main gaps are illustrative rather than real APIs in the primary examples and a missing explicit validate-retry loop on the main data path.

Suggestions

Replace the illustrative read_file()/has_reference_file()/search_web() pseudocode with real, copy-paste-ready tool calls so the primary path is fully executable.

Tighten the intro and 'Core Principle' sections to avoid restating the description, and remove the decision-flow pseudocode in favor of a one-line rule.

Add an explicit validate-and-retry checkpoint in Step 3 (e.g., 'If extracted data is incomplete or unreadable, re-read with a fallback method before proceeding').

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient with concrete code, but the intro paragraph and 'Core Principle' restate what the description already says, and some explanatory phrasing could be trimmed. It is above 2 (no extensive concept explanations) but below 4 (minor over-explanation remains).	3 / 5
Actionability	The DOCX fallback provides genuinely executable bash and python snippets, but the primary read_file()/has_reference_file()/search_web() calls are illustrative non-real APIs rather than copy-paste-ready code. It is above 3 (real commands present) but below 5 (not fully executable throughout).	4 / 5
Workflow Clarity	A clear 5-step sequence (Scan, Read, Extract/Validate, Use as primary, Search web when necessary) is given, with a verify step in the DOCX fallback and a closing checklist. It is above 3 (checkpoints present) but below 5 (the main data-extraction path lacks an explicit validate-and-retry feedback loop).	4 / 5
Progressive Disclosure	The skill is a single self-contained file with well-organized sections and no bundle files to navigate, so structure is good. It is above 3 (clear sectioning, not a wall of text) but below 5 because the all-inline DOCX troubleshooting could arguably live in a reference, leaving minor organization room.	4 / 5
	Total	15 / 20 Passed

Description

38%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.

The description conveys a clear behavioral directive but lacks natural trigger terms and an explicit 'Use when...' clause, making it unlikely to be selected on its own. It is also too generic to distinguish itself from other data-handling skills.

Suggestions

Add an explicit 'Use when...' clause naming concrete trigger phrases users would say (e.g., 'Use when the user provides data files such as .xlsx, .csv, .pdf, or .docx and asks for analysis or output generation').

Replace generic terms ('reference files', 'data') with concrete file extensions and natural synonyms users actually mention.

Narrow the niche to reduce overlap, e.g., 'Reads provided spreadsheet, CSV, PDF, and DOCX attachments and uses their contents as the primary data source for outputs.'

Dimension	Reasoning	Score
Specificity	Names the domain (reference files) and 1-2 concrete actions ('read and use provided reference files for data', 'before attempting external searches or fabricating information'), but coverage is not comprehensive. It is below 4 (which requires several specific actions) and above 2 (which only names the domain with minimal actions).	3 / 5
Completeness	It has a clear 'what' (read and use provided reference files for data) but no 'Use when...' clause or explicit trigger guidance, capping completeness at 3 per the rubric guidelines. It is above 2 (vague what with no when) but below 4 (which requires both what and when).	3 / 5
Trigger Term Quality	Keywords ('reference files', 'data', 'external searches', 'fabricating information') are generic/technical rather than natural phrases a user would actually say. It is above 1 because some relevant terms exist, but below 3 since common natural variations and synonyms are missing.	2 / 5
Distinctiveness Conflict Risk	It describes a broad meta-behavior applicable to virtually any data-handling task, creating high overlap risk with many skills. It is above 1 (not entirely empty of specificity) but below 3 (still very broad rather than 'somewhat specific').	2 / 5
	Total	10 / 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: HKUDS/OpenSpace
Path: benchmarks/gdpval/skills/prioritize-reference-files-6e4111/SKILL.md
Commit: 3827781

Reviewed: 9 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.