cluster-agent-swarm

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs agents to fetch and inspect artefacts and metadata from external registries and services (e.g., skills/artifacts/SKILL.md shows crane/skopeo/curl/jfrog/Harbor/ACR/ECR commands and trivy/syft/cosign SBOM and scan outputs) — these are untrusted, user-provided third‑party contents that the agents must read and which directly gate decisions such as artifact promotion and escalation, enabling indirect prompt‑injection risk.