linkerd-patterns

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow includes fetching and ingesting the contents of an outsider-authored installer script from https://run.linkerd.io/install (via curl ... -o ... and cat "$tmpdir/linkerd-install.sh"), which becomes readable free text in the agent context before execution.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill includes a runtime command that downloads and executes a remote installer script from https://run.linkerd.io/install (curl ... -o ...; sh "$tmpdir/linkerd-install.sh"), which fetches and executes remote code during runtime.