create-workflow

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

• Third-party content exposure detected (medium risk: 0.65). The workflow writer skill composes Nextflow workflows by including modules from the Nextflow Registry, and at runtime the run-module validation path uses nextflow module search/view and nextflow module run which fetches outsider-authored module metadata/code from the public Nextflow Registry and injects it into the agent’s LLM context via the tool’s readable command outputs/logs.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill uses Nextflow module commands to download and run modules from the Nextflow Registry at runtime (e.g., https://registry.nextflow.io), which fetches remote code that is then executed as part of the workflow.