omnidocbench-eval-helper

Help users deploy, validate, run, and parse OmniDocBench evaluations. Use this skill whenever the user mentions OmniDocBench, document parsing/OCR benchmark scoring, MinerU or other model evaluation on OmniDocBench, CDM formula metrics, end2end/md2md configs, Docker/conda deployment, remote SSH/H-cluster execution, result JSON parsing, or troubleshooting TeX Live/ImageMagick/Ghostscript/Docker/worker/OOM issues. Prefer Docker first, generate concrete commands from the user's paths, validate inputs before running, and report final Overall/Text/Formula/Table/Reading-order scores with result file paths.

Quality

92%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Medium

Suggest reviewing before use

Quality

Content

85%

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

The content is highly actionable with executable commands, a well-sequenced workflow including validation checkpoints, and good progressive disclosure via two real bundled scripts. The main weakness is conciseness: the large duplicated end2end YAML block and repeated worker guidance inflate the token budget where the provided config script already handles generation.

Suggestions

Replace the verbatim YAML heredoc inside both the local and remote Docker commands with a call to scripts/generate_end2end_config.py (already bundled) to remove a large duplicated block and the redundant worker tuning notes.

Consolidate the repeated '4CPU/8G use 2, use 1 if unstable' worker guidance into a single statement in Key facts and reference it elsewhere, rather than restating it in three places.

Consider splitting the long troubleshooting playbook and result-parsing details into a separate reference file referenced one level deep, to keep SKILL.md a leaner overview.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient and full of directly useful, executable content, but it is long (~400 lines) with some redundancy (the end2end YAML block is duplicated verbatim between the local and remote sections, and worker guidance is restated several times), so it could be tightened by referencing the config-generating script instead of repeating YAML.	2 / 3
Actionability	Provides fully executable, copy-paste-ready bash commands for validation, Docker runs, remote SSH, conda, and a concrete result-field extraction block plus bundled scripts with exact invocations and flags, matching the level-3 anchor.	3 / 3
Workflow Clarity	A clearly numbered top-level sequence (validate paths -> generate config -> run in Docker -> parse result files -> report scores) with explicit validation checkpoints before launching long jobs and an error-recovery-oriented troubleshooting playbook, satisfying the level-3 anchor with feedback loops for risky/batch operations.	3 / 3
Progressive Disclosure	The body is organized into clear sections and offloads real detail to two actual bundled scripts (generate_end2end_config.py, parse_results.py) that are referenced one level deep with concrete usage, rather than nesting references or dumping everything inline, matching the level-3 anchor.	3 / 3
	Total	11 / 12 Passed

Description

100%

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 is specific, trigger-rich, complete, and distinctive, explicitly covering both what the skill does and when to use it with concrete natural-language trigger terms. It is on the longer side but every clause adds a concrete capability or trigger rather than fluff.

Dimension	Reasoning	Score
Specificity	Lists many concrete actions ('deploy, validate, run, and parse OmniDocBench evaluations', 'generate concrete commands from the user's paths', 'report final Overall/Text/Formula/Table/Reading-order scores'), matching the level-3 anchor of multiple specific concrete actions.	3 / 3
Completeness	Explicitly states both what it does (deploy, validate, run, parse, report scores) and when to use it ('Use this skill whenever the user mentions OmniDocBench...'), satisfying the level-3 anchor for both what and when.	3 / 3
Trigger Term Quality	Strong coverage of natural user terms (OmniDocBench, MinerU, CDM formula metrics, end2end/md2md configs, Docker/conda deployment, remote SSH/H-cluster, TeX Live/ImageMagick/Ghostscript, OOM) that a user would actually say, matching the level-3 anchor.	3 / 3
Distinctiveness Conflict Risk	A clear niche (OmniDocBench benchmark evaluation) with distinct, domain-specific triggers unlikely to conflict with other skills, matching the level-3 anchor.	3 / 3
	Total	12 / 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: opendatalab/OmniDocBench
Path: skills/SKILL.md
Commit: 193627a

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.