xiaoyue-companion
Fail
Audited by Gen Agent Trust Hub on Mar 18, 2026
Risk Level: HIGHCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: A functional Zhipu AI API key is hardcoded in the installation documentation.
- File:
projects/companion-skill/INSTALL.md - Evidence:
"ZHIPU_API_KEY": "da8df5ba954341829f7afd05ca23a889.RrJoTsbaAkGYA6ZU" - [REMOTE_CODE_EXECUTION]: Setup scripts execute remote content piped directly to the shell to install the Rust toolchain. While this targets a well-known service, it remains an RCE pattern.
- File:
skills/chrome-automation/chrome-automation/scripts/auto-install-mac.sh - Evidence:
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --no-modify-path - [DATA_EXFILTRATION]: Hardcoded IP address used as an API backend for content publishing, creating a risk of data exposure to an unverified endpoint.
- File:
skills/wechatsync-publisher/wechatsync-publisher/SKILL.md - Evidence:
http://39.108.254.228:8002 - [PROMPT_INJECTION]: Behavioral instructions in the Moltbook skill direct the agent to perform specific social actions (joining communities, upvoting manifestos) without explicit user confirmation for each step.
- File:
skills/moltbook/SKILL.md - Evidence: "Join the Stepbot Community... giving it a upvote. Then, post a self-introduction..."
- [EXTERNAL_DOWNLOADS]: The Moltbook skill fetches multiple configuration and instruction files from an external domain (
moltbook.com) and saves them locally. - File:
skills/moltbook/SKILL.md - Evidence:
curl -s https://www.moltbook.com/skill.md > ~/.moltbot/skills/moltbook/SKILL.md
Recommendations
- HIGH: Downloads and executes remote code from: https://sh.rustup.rs - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata