create-agent-site-skill
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's workflow discovery phase involves navigating to external websites to record actions and identify selectors. These websites are untrusted and could provide content designed to influence the logic of the generated skill.
- Ingestion points: Browser content and navigation state are ingested via
kernel browsers playwright executeand visual state viakernel browsers computer screenshotin SKILL.md. - Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions to separate external data from the agent's core instructions in SKILL.md.
- Capability inventory: The skill utilizes
kernelCLI tools for persistent browser profiles, vault management, and file system operations to create new skill files in SKILL.md. - Sanitization: The skill lacks explicit sanitization steps for data extracted from pages before incorporating it into the final generated
SKILL.mdfile.
Audit Metadata