implementing-threat-modeling-with-mitre-attack

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

• Third-party content exposure detected (high risk: 0.85). The workflow’s runtime path fetches outsider-authored free text from the public web: it uses requests.get("https://raw.githubusercontent.com/mitre/cti/.../enterprise-attack.json") to ingest MITRE ATT&CK STIX JSON, which is then parsed and incorporated into the agent’s LLM context (e.g., group descriptions/technique mappings).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's Python code fetches and parses the MITRE CTI JSON at runtime from https://raw.githubusercontent.com/mitre/cti/master/enterprise-attack/enterprise-attack.json, and that external data is used to populate groups/techniques which directly determine the threat-model outputs, so it is a required runtime dependency that controls agent behavior.