website-source-for-agents
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to read external, potentially untrusted content such as a project's README, editing guide, and setup prompt to perform its tasks. This represents a vulnerability surface where a malicious project could embed instructions to manipulate the agent's behavior.
- Ingestion points: Project files, specifically README, editing guides, and setup prompts located in the user-downloaded folder (SKILL.md).
- Boundary markers: None specified; the agent is instructed to follow the instructions found in the files directly.
- Capability inventory: The agent is empowered to modify source code, change project configuration, and execute "documented preview commands" (SKILL.md).
- Sanitization: No sanitization or validation of the external instructions is mentioned.
- [EXTERNAL_DOWNLOADS]: The skill directs users to external URLs (getplate.pages.dev) to download website source code. While the user performs the download, the agent is subsequently instructed to process and execute content from these external packages.
Audit Metadata