galaxy-workflow-development
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acts as an expert in processing and validating Galaxy workflow files (.ga), which are JSON-based data structures containing various natural-language metadata fields. These fields, such as annotations and documentation blocks, could be used to inject malicious instructions into the agent's context.\n
- Ingestion points: Processes workflow JSON (.ga), Planemo test YAML, and repository metadata files.\n
- Capability inventory: Access to system shell for running Planemo, file I/O for editing JSON, and network access for interacting with Galaxy APIs.\n
- Boundary markers: No specific delimiters or filtering rules are implemented to isolate user-provided text in the workflows from agent instructions.\n
- Sanitization: Lacks specific sanitization logic for content extracted from workflow files.\n- [DYNAMIC_EXECUTION]: The skill provides numerous Python and shell script templates to automate metadata updates and testing verification. These templates generate code that is executed locally to process and modify workflow files.\n
- Evidence: Multiple Python heredocs in iwc-standards.md and testing-guide.md for tasks like cleaning runtime parameters and comparing tool versions.\n
- Pattern: Uses script generation from known templates to process local data files.
Audit Metadata