website-copy
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill does not contain any executable scripts such as Python, JavaScript, or Shell scripts. It is composed entirely of Markdown instructions, JSON evaluation data, and YAML configuration.
- [INDIRECT_PROMPT_INJECTION]: The skill instructions (Step 3 in
SKILL.md) direct the agent to fetch and read content from existing websites using tools likeagent-browser. This creates a potential ingestion point for indirect prompt injection if a fetched site contains malicious instructions. However, the risk is minimal as the skill lacks dangerous capabilities like arbitrary code execution or file system access. - Ingestion points: External web content fetched via
agent-browserorfetchas described inSKILL.md. - Boundary markers: Absent.
- Capability inventory: No executable scripts, subprocess calls, or system-level tools are present in the skill files.
- Sanitization: No explicit sanitization of fetched content is described in the instructions.
Audit Metadata