The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute nextflow CLI commands. These commands are used for module management (search, view, run) and executing the final workflow script.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download of bioinformatics modules and container images from the Nextflow Registry (registry.nextflow.io). This is the standard, documented behavior for Nextflow development and uses the vendor's official infrastructure.
[REMOTE_CODE_EXECUTION]: The skill executes code downloaded from the Nextflow Registry and dynamically generates workflow scripts. This behavior is the primary purpose of the skill and is mitigated by a mandatory 4-step process requiring individual module validation and explicit user approval before final composition.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from the Nextflow Registry and user-provided inputs.
Ingestion points: Output from nextflow module search and nextflow module view (SKILL.md).
Boundary markers: The instructions use structured steps and wait for user approval, but do not use explicit delimiters for tool outputs.
Capability inventory: The agent has access to Bash for command execution, Write for file creation, and Skill for delegating to other tasks (SKILL.md).
Sanitization: The skill enforces a mandatory user approval step (Step 2) and a sequential validation step (Step 3) before any composite workflow is executed, which serves as a significant manual checkpoint.

create-workflow