sdlc
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [No Code] (SAFE): The skill consists entirely of markdown documentation and process instructions. No executable scripts, binaries, or configuration files are present.
- [Indirect Prompt Injection] (LOW): The skill defines a surface for indirect injection as it processes external code through multiple subagent roles without explicit boundary markers. 1. Ingestion points: External code provided for review (SKILL.md). 2. Boundary markers: Absent; no specific instructions to ignore embedded commands in reviewed code. 3. Capability inventory: Subagent spawning and code refactoring capabilities across multiple scripts. 4. Sanitization: Absent; no validation or escaping of the code provided for review.
Audit Metadata