blog-outline
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted data from the web.
- Ingestion points: In Step 2, the skill uses the
WebFetchtool to extract detailed content and heading structures from external competitor websites identified via search results. - Boundary markers: The instructions lack explicit delimiters or warnings to the agent to ignore any natural language instructions that might be embedded within the fetched competitor content.
- Capability inventory: The skill has access to powerful tools including
BashandWrite, which could be exploited if the agent follows malicious instructions hidden in external web pages (e.g., instructions to delete files or execute shell commands). - Sanitization: There are no requirements or steps provided to sanitize, filter, or validate the content retrieved from external URLs before the agent processes it for outline generation.
Audit Metadata