Back to skill
Skillv1.0.3
ClawScan security
AGENT-OS · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMay 1, 2026, 6:36 AM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5.5
- Summary
- This instruction-only skill does not request credentials, tools, file access, code execution, persistence, or external APIs.
- Guidance
- The provided artifacts show a minimal instruction-only skill with no requested permissions or executable components. Its purpose is broad, so users should understand that it is described as an agent-routing layer, but there is no evidence here of risky access or hidden behavior.
Review Dimensions
- Purpose & Capability
- okThe stated purpose is broad agent coordination, but the provided artifact contains only a high-level description and no concrete instructions that expand authority or change user data.
- Instruction Scope
- okThe SKILL.md content does not include goal-overriding instructions, forced tool use, unsafe defaults, or autonomous high-impact actions.
- Install Mechanism
- okThere is no install spec, no code files, no required binaries, and no package or script execution.
- Credentials
- okThe registry declares no required environment variables, credentials, config paths, or OS-specific permissions.
- Persistence & Privilege
- okThe artifacts show no persistence, background behavior, credential use, local indexing, memory storage, or privilege escalation.