Back to skill
Skillv1.0.0
ClawScan security
职场高情商沟通润色/职场防背锅助手 · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMay 1, 2026, 7:38 AM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5.5
- Summary
- This is an instruction-only workplace communication polishing skill with no code, credentials, tools, persistence, or hidden install behavior.
- Guidance
- Before sending any generated workplace message, review it for factual accuracy, tone, and whether it appropriately reflects your responsibility boundaries and company context.
Review Dimensions
- Purpose & Capability
- okThe stated purpose—rewriting emotionally charged workplace messages into professional, goal-oriented drafts—is consistent with the SKILL.md workflow and constraints.
- Instruction Scope
- okInstructions are limited to analyzing user-provided drafts and producing copyable communication options; they do not ask the agent to send messages, access accounts, override user intent, or perform external actions.
- Install Mechanism
- okThere is no install spec, no code, no binaries, no package dependencies, and no setup commands.
- Credentials
- okThe skill requests no environment variables, credentials, local files, network access, or system capabilities.
- Persistence & Privilege
- okThe artifacts show no persistence, background behavior, privilege escalation, memory use, credential handling, or account mutation authority.