sealos-deploy

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly clones and reads arbitrary GitHub repositories (modules/preflight.md Step 2.1 and modules/pipeline.md Phase 1), scans README/.github workflows and external registries (modules/pipeline.md Phase 2) and uses those untrusted, public contents to drive detection, build, registry selection, and deployment decisions — allowing indirect instruction-like content to influence tool use and next actions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill includes a runtime installation step that runs a remote shell script via curl -fsSL https://get.docker.com | sh (used when installing Docker during preflight if the user agrees), which fetches and executes remote code at runtime and can be required to enable the build path—this meets the criteria for a high-confidence risky external dependency.