building-ioc-enrichment-pipeline-with-opencti

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

• Third-party content exposure detected (high risk: 1.00). The skill's code and workflow explicitly call public third-party APIs (e.g., scripts/process.py and SKILL.md showing VirusTotal, Shodan, AbuseIPDB, GreyNoise, SecurityTrails calls) and directly ingest their JSON responses to build notes, set labels, and compute/update confidence scores in OpenCTI—meaning untrusted/user-contributed content (notably AbuseIPDB reports and returned fields/tags) is read and can materially drive automated decisions and tool actions.

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

• Potentially malicious external URL detected (high risk: 0.70). The docker-compose references (e.g., image: opencti/platform:6.4.4 and image: opencti/connector-virustotal:6.4.4 — effectively docker.io/opencti/platform:6.4.4 and docker.io/opencti/connector-virustotal:6.4.4) are fetched at deployment/runtime and will execute remote code as required components of the skill, so they meet the criteria for a runtime external dependency that can execute code.