prioritize-reference-files

Ensures agents read and use provided reference files before searching or fabricating data

Quality

56%

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

Quality

Content

62%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is a well-sequenced, actionable workflow with real fallback commands and validation checkpoints, scoring high on workflow clarity. Its weaknesses are an invented read_file API that undercuts actionability, minor verbosity, and a monolithic structure with no external reference files despite substantial inline .docx detail.

Suggestions

Replace the invented read_file(filetype=..., file_path=...) calls with the actual available tool invocations, or explicitly note the abstraction, so examples are copy-paste executable.

Move the lengthy .docx fallback extraction into a references/DOCX_FALLBACK.md and link to it, leaving the main flow lean.

Tighten Step 3 and Step 5 and trim the anti-patterns list, which largely restates the Core Principle, to improve token efficiency.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient with actionable structure, but Step 3 ('Parse the content to understand the data structure'), Step 5 source-documentation boilerplate, and the anti-patterns repeat guidance Claude can infer, so it is anchor 2 rather than the lean anchor 3.	2 / 3
Actionability	Provides genuinely executable bash fallbacks (unzip/grep of word/document.xml) but the primary read_file(filetype=..., file_path=...) calls are an invented, non-standard tool signature rather than copy-paste executable code, landing at anchor 2.	2 / 3
Workflow Clarity	Five steps are clearly sequenced, with explicit validation in Step 3 and Fallback Approach 3 ('Verify extraction success'), feedback loops ('try an alternative method'), and a verification checklist, matching anchor 3's clear-sequence-with-checkpoints bar.	3 / 3
Progressive Disclosure	Sections are well organized, but the file is over 50 lines with no external references and substantial inline detail (the .docx fallback extraction) that could be split into a reference file, so it is anchor 2 rather than the cleanly-split anchor 3.	2 / 3
	Total	9 / 12 Passed

Description

50%

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 the skill's purpose and domain but is generic, lacks an explicit 'Use when' trigger clause, and uses some non-natural jargon like 'fabricating data'. It is distinguishable but not sharply differentiated from general file-handling skills.

Suggestions

Add an explicit trigger clause, e.g. 'Use when reference files (spreadsheets, PDFs, CSVs, JSON) are provided in task context or when the user references an attached or provided file.'

Replace jargon like 'fabricating data' with natural user phrasings such as 'making up data' or 'inventing data'.

List concrete file formats and actions to sharpen specificity, e.g. 'Reads .xlsx, .csv, .pdf, .json, .docx files and extracts their data before searching the web.'

Dimension	Reasoning	Score
Specificity	Names the domain ('reference files') and some actions ('read and use', 'before searching or fabricating data'), but does not list multiple concrete actions, matching anchor 2 rather than the multi-action anchor 3.	2 / 3
Completeness	It states what the skill does but lacks any 'Use when...' or equivalent explicit trigger clause, which per the guidelines caps completeness at 2 rather than 3.	2 / 3
Trigger Term Quality	Contains relevant keywords like 'reference files' but 'fabricating data' is jargon users would not naturally say, and common variations ('attached file', 'the spreadsheet I provided') are missing, so it sits at anchor 2 rather than 1 or 3.	2 / 3
Distinctiveness Conflict Risk	'Reference files' is a somewhat specific niche but the phrasing is generic enough to overlap with general file-handling skills, fitting anchor 2 rather than the clearly-distinct anchor 3.	2 / 3
	Total	8 / 12 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/SKILL.md
Commit: c5a9c4b

Reviewed: 2 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.