Back to skill
Skillv1.0.0
ClawScan security
Arch Review · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMay 1, 2026, 8:30 AM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5.5
- Summary
- This is an instruction-only architecture review aid with no code, install steps, credentials, persistence, or system access.
- Guidance
- This skill appears safe to install as an advisory prompt. As with any architecture review, avoid pasting real secrets, private keys, or unnecessary confidential details into the conversation.
Review Dimensions
- Purpose & Capability
- okThe stated purpose is to help stress-test architecture designs, surface risks, and produce review notes or ADR-style outputs; the provided instructions are aligned with that advisory purpose.
- Instruction Scope
- okThe skill asks for architecture context and directs the agent to summarize, identify risks, mitigations, and open questions; it does not instruct the agent to override user goals, force tool use, or perform high-impact actions.
- Install Mechanism
- okThere is no install specification, no code files, no required binaries, and no required environment variables.
- Credentials
- okThe artifacts do not request local file access, network access, credentials, API access, shell commands, or broad environmental authority.
- Persistence & Privilege
- okThe artifacts show no persistence, background activity, privilege escalation, memory storage, or autonomous behavior outside the review interaction.