Skillv0.1.3

ClawScan security

AI纠错 · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

BenignApr 6, 2026, 9:48 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: This skill is an instruction-only Chinese alias that maps user intents (like “模型纠偏”) to the official VeriClaw skill; it requests no credentials, installs, or system access and is internally coherent with its stated purpose.
Guidance: This skill is low-risk and simply maps several Chinese correction-related intents to the VeriClaw skill. If you want the official, full-featured installation, follow the SKILL.md advice and install 'vericlaw' from the canonical source (or the listed homepage). Because it's instruction-only and asks for no credentials or installs, there's no obvious security concern here — but be aware this alias may not implement functionality itself; installing the main vericlaw skill is required for full capabilities. As a general precaution, only install the canonical skill from the official repo/homepage listed if you need the full toolset.

Purpose & Capability: okName and description say this is an alias/route into VeriClaw; the skill requires no binaries, env vars, or install steps—consistent with being a lightweight routing/intent-mapping skill.
Instruction Scope: okSKILL.md only documents routing rules, canonical pages, and user guidance (e.g., tell users to install 'vericlaw' for the canonical surface). It does not instruct reading files, accessing credentials, or calling unrelated external endpoints.
Install Mechanism: okNo install spec or code files are present; instruction-only skills have minimal footprint and nothing is written to disk or fetched during install.
Credentials: okNo environment variables, credentials, or config paths are requested; proportional to the described intent-mapping role.
Persistence & Privilege: okFlags are default (always: false, autonomous invocation allowed). The skill does not request persistent system privileges or modify other skills' configs.