Software Factory Workflow
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides documentation and process guidelines for a five-agent software development lifecycle. No functional scripts or tool configurations are included in the file. All code snippets are provided as static examples for illustrative purposes.
- [INDIRECT_PROMPT_INJECTION]: The methodology establishes a data flow where raw user ideas from the 'project-idea/' directory are processed by agents. This pattern constitutes a potential surface for indirect prompt injection attacks. * Ingestion points: 'project-idea/index.md' and related 'amendments/' files. * Boundary markers: No explicit instructions for delimiters or instruction-bypass protection are included in this high-level workflow overview. * Capability inventory: The 'Assembler' agent's role includes writing code and executing test suites. * Sanitization: The workflow does not detail input validation or sanitization procedures for the initial idea phase.
Audit Metadata