pipeline-orchestrator
Pass
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes git commands to manage development workflows, including creating worktrees and pushing or deleting branches. These operations are restricted to a feature-specific scope and are documented to include user confirmation steps for destructive actions.
- [PROMPT_INJECTION]: The skill processes untrusted input from external kanban boards and story descriptions (Indirect Prompt Injection surface).
- Ingestion points: Story descriptions and task lists parsed during Phase 1.
- Boundary markers: None explicitly defined for the external data.
- Capability inventory: File system writes (.pipeline/state.json), git command execution, and delegation to multiple sub-skills (coordinators).
- Sanitization: Relies on a 'multi-agent-validator' stage to verify plans before execution, providing a layer of logic validation.
Audit Metadata