stage
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, creating an attack surface where embedded instructions in the input could influence the agent's behavior.
- Ingestion points: Untrusted data is ingested from agents like Scribe, Tome, and Canvas, as well as from user-provided outlines and topics defined in the workflow.
- Boundary markers: The skill instructions do not specify the use of delimiters or cautionary instructions to the model to ignore potentially malicious directions within the ingested content.
- Capability inventory: The skill produces executable Markdown and HTML slide files and provides CLI instructions for previewing them.
- Sanitization: There is no evidence of content sanitization or escaping mechanisms to prevent prompt injection from being carried over into the generated slide code.
- [SAFE]: The skill is primarily focused on legitimate presentation drafting and uses verified resources.
- External References: All framework documentation links point to official and reputable domains (e.g., marp.app, sli.dev, revealjs.com).
- Standard Tooling: The suggested preview methodology utilizes the official @marp-team/marp-cli package, which is a standard community tool.
Audit Metadata