product-box-block
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches product details from external, untrusted URLs using tools like WebFetch or Firecrawl. This creates a surface where an attacker-controlled website could include malicious instructions in product titles or descriptions to influence the agent's behavior.
- Ingestion points: Product details are fetched from external URLs provided by the user or clipboard (SKILL.md).
- Boundary markers: There are no instructions to the agent to use delimiters or to ignore instructions embedded in the fetched content.
- Capability inventory: The skill has the ability to execute shell commands (pbcopy) and read from the clipboard (SKILL.md).
- Sanitization: The instructions do not specify any validation or sanitization for the data extracted from web pages before it is used in the output or shell commands.
- [COMMAND_EXECUTION]: The skill explicitly instructs the agent to execute a shell command (
echo '...' | pbcopy) to transfer the generated markup to the system clipboard. Although this is the primary function of the skill, piping data derived from external websites directly into a shell execution environment is a potential risk factor if the agent does not correctly escape the content.
Audit Metadata