parallel-task
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes markdown plan files to extract task descriptions, acceptance criteria, and validation commands which are then directly interpolated into prompts for subagents. A malicious plan file could contain instructions designed to override the subagent's safety constraints or execute unintended actions.
- Ingestion points: Markdown plan files parsed in Step 2 and Step 3.
- Boundary markers: The skill does not define explicit delimiters or instructions for subagents to ignore potential command injections within the task content extracted from the plan.
- Capability inventory: Subagents are instructed to perform codebase research, modify production files, execute shell commands for validation (e.g., TDD test runs), and perform git commits.
- Sanitization: There is no evidence of sanitization, escaping, or validation of the text extracted from the plan files before it is used to form the subagent prompt template.
Audit Metadata