newsletter-coach
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- General Analysis (SAFE): The provided files consist of documentation (GUIDE.md, ROADMAP.md) and resource templates (questions, outlines, and writing guides). No executable code, installation scripts, or system configuration files were present.
- Indirect Prompt Injection (LOW): The skill's primary function is to process user-provided stories and ideas into structured content. While this represents an indirect prompt injection surface (processing untrusted input), it is the core intended use case and lacks dangerous capabilities like shell access or data exfiltration that would escalate the risk.
- Ingestion points: User input provided during the 7-phase coaching process (Stop 1).
- Boundary markers: Not explicitly defined in resource files; likely handled by the main agent instructions.
- Capability inventory: Text generation (drafting), headline generation, and potential image generation (referenced in roadmap).
- Sanitization: Not present in the static resource templates.
- External Dependencies (SAFE): No package managers (pip, npm) or remote script downloads were found. The roadmap mentions 'nano-banana-pro' for image generation, which appears to be an internal tool call rather than an external execution risk.
- Data Exposure (SAFE): No hardcoded credentials, API keys, or sensitive local file paths were detected.
Audit Metadata