Skillv1.0.0

ClawScan security

顶级名医中医辨证Skill · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

SuspiciousApr 8, 2026, 5:16 AM

Verdict: suspicious
Confidence: high
Model: gpt-5-mini
Summary: The skill's description promises a full interactive TCM diagnostic and prescription workflow, but the included code is trivial (only prints two lines) and does not implement the claimed functionality—this mismatch is concerning.
Guidance: The skill's documentation promises a full-featured TCM diagnostic system, but the included script is a stub that only prints two lines and does not perform the claimed问诊、辨证、开方流程. Treat the package as incomplete or misleading. Before installing or using it for any health-related purpose: (1) ask the publisher for the complete implementation or provenance (homepage/source repo), (2) inspect any additional code that might be fetched at runtime, (3) run the code in a safe, isolated environment if you want to test it, and (4) do not rely on it for medical advice—it explicitly states it is for education and the implementation does not match the advertised capability.

Purpose & Capability: concernThe name/description advertise a comprehensive intelligent问诊→辨证→开方 system, which would legitimately require data input, diagnostic logic, and possibly external knowledge sources. The skill only declares python3 as a requirement and ships a tiny script that does not implement those capabilities, indicating the declared purpose does not match the actual capability.
Instruction Scope: concernSKILL.md instructs users to run 'python3 scripts/tcm.py' and implies the program will automatically complete the full diagnostic and prescribing workflow. The actual script merely prints two informational lines and contains static data structures—there is no interactive question flow, no processing of user input, and no generation of prescriptions as described. The instructions therefore overstate what will happen and grant the agent broad discretion that the code does not execute.
Install Mechanism: okNo install spec is present (instruction-only), and the only runtime dependency is python3. This is proportionate and low risk from an install perspective.
Credentials: okThe skill requests no environment variables, no credentials, and no config paths. Its minimal requirements are proportionate to the simple code provided.
Persistence & Privilege: okalways is false and the skill does not request persistent presence or elevated privileges. There is no evidence it modifies other skills or agent-wide settings.