btst-ai-context
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents patterns for passing page-level data into an AI context, which creates an attack surface for indirect prompt injection.
- Ingestion points: The
pageDescriptionfield in theuseRegisterPageAIContexthook (SKILL.md) is used to ingest dynamic page content, such as titles and body text, into the AI's context. - Boundary markers: The code snippets demonstrate direct interpolation of page content into the description string without the use of clear delimiters or instructions to ignore embedded commands.
- Capability inventory: The skill implements
clientTools(SKILL.md) which provide the AI with the ability to perform actions on the user's behalf, such as usingformRef.current.setValueto modify form state. - Sanitization: The documentation does not include examples or requirements for sanitizing or validating the page content before it is processed by the AI widget.
Audit Metadata