laohan-weigui

Pass

Audited by Gen Agent Trust Hub on Jul 25, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a Node.js script located at scripts/scan.mjs to perform keyword scanning. This script reads from the local file system (specifically 01-口播稿.md) and writes analysis results to a JSON file (02-违规扫描.json). The script uses standard built-in modules (node:fs, node:crypto, node:path) and does not invoke external shell commands or untrusted binaries.
  • [PROMPT_INJECTION]: While the skill processes user-supplied text (scripts for scanning), which is a common surface for indirect prompt injection, it mitigates this by using a structured workflow. The processing logic is focused on keyword matching and heuristic evaluation rather than executing the instructions contained within the user text.
  • [DATA_EXFILTRATION]: There are no network operations detected in the instructions or the accompanying JavaScript tool. All data processing and report generation occur locally within the workspace.
  • [REMOTE_CODE_EXECUTION]: No external package installations or remote script downloads are performed. All logic is contained within the provided skill files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 25, 2026, 03:31 PM
Security Audit — agent-trust-hub — laohan-weigui