verified-frontend
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill instructs the agent to analyze local repository data (frameworks, start commands, and CI configurations) to build a project-specific workflow. This creates a surface for indirect prompt injection if an attacker places malicious instructions inside the files the agent is directed to inspect.
- Ingestion points: Repository inspection steps defined in
SKILL.mdfor discovering project structure and tools. - Boundary markers: None are specified to delimit ingested repository data from instructions.
- Capability inventory: The instructions involve running local applications, modifying source code ("Fix visible problems"), and capturing system screenshots.
- Sanitization: No sanitization or validation of the project-sourced data is performed before it influences the agent's task generation.
Audit Metadata