mitre-attack-lookup

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). This skill explicitly clones and reads files from the public mitrize GitHub repository (https://github.com/woohyun212/mitrize) and instructs the agent to cat/grep/query those Markdown files as part of its lookup workflow, so untrusted third‑party content is fetched and interpreted in ways that can influence decisions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly clones and then runs code from the remote repository https://github.com/woohyun212/mitrize.git (git clone ... then python3 scripts/query_attack_md.py), so fetched external content is executed at runtime and is a required dependency.