Back to skill
Skillv1.0.0
ClawScan security
PaperRead-Skills · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMay 1, 2026, 6:13 AM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5.5
- Summary
- This instruction-only skill is narrowly focused on analyzing a user-uploaded paper PDF and does not request credentials, installation, code execution, persistence, or external access.
- Guidance
- This appears safe to install for its stated purpose. Use it when you want an evidence-grounded critique of a paper PDF, and remember that it will base its output on the uploaded document rather than outside sources.
Review Dimensions
- Purpose & Capability
- okThe stated purpose and included prompts are coherent: extract facts from an uploaded CVPR PDF, then produce an evidence-grounded critique.
- Instruction Scope
- okThe instructions constrain the agent to the uploaded PDF, require evidence tags, and tell it to state when information is missing or insufficient.
- Install Mechanism
- okThere is no install spec, no required binaries, no required environment variables, and no code files.
- Credentials
- okThe skill does not request filesystem, network, credential, shell, or account access beyond reading the user-provided PDF in the current chat.
- Persistence & Privilege
- okNo persistence, background process, privilege escalation, credential use, or stored memory behavior is described in the artifacts.