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.
75
92%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
Medium
Suggest reviewing before use
Security
1 medium severity finding. This skill can be installed but you should review these findings before use.
W013: Attempt to modify system services in skill instructions.
What this means
The skill prompts the agent to compromise the security or integrity of the user’s machine by modifying system-level services or configurations, such as obtaining elevated privileges, altering startup scripts, or changing system-wide settings.
Why it was flagged
The skill explicitly instructs use of sudo (e.g., `sudo -n docker`) and advises modifying system configuration files (adjusting ImageMagick `policy.xml`) and installing system-wide packages (TeX Live/CJK), which require elevated privileges and can change the host state.
Low
Low-risk findings.
2 low severity findings. Worth noting, but not necessarily harmful.
W011: Third-party content exposure detected (indirect prompt injection risk).
What this means
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
Why it was flagged
The required runtime workflow ingests outsider-authored free text via the user-provided prediction markdown directory (and possibly ground-truth JSON) mounted into the Docker evaluation, which the LLM backend presumably uses to compute metrics (e.g., `-v "$PRED":.../predictions` then `python pdf_validation.py --config ...`).
W012: Unverifiable external dependency detected (runtime URL that controls agent).
What this means
The skill fetches instructions or code from an external URL at runtime, and the fetched content directly controls the agent’s prompts or executes code. This dynamic dependency allows the external source to modify the agent’s behavior without any changes to the skill itself.
Why it was flagged
The skill pulls and runs the remote Docker image ghcr.io/zeng-weijun/omnidocbench-eval:repro-ubuntu2204 at runtime (docker pull / docker run), which fetches and executes remote code the skill relies on.
- Repository
- opendatalab/OmniDocBench
- Path
skills/SKILL.md
- Commit
193627a
- Audited
- Security analysis
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.