moda-deck-pptx
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external
.pptxfiles, which may contain maliciously crafted content designed to influence the agent's instructions during the analysis and fix stages. - Ingestion points: PowerPoint files imported via
moda canvas import-pptx(SKILL.md). - Boundary markers: No explicit delimiters or warnings are defined to isolate untrusted file data from the agent's primary instructions.
- Capability inventory: The agent has access to the
Bash(moda:*)toolset, allowing it to read/write canvases, export files, and interact with media/web subcommands (SKILL.md, references/omni-and-media.md). - Sanitization: There is no evidence of content sanitization or filtering applied to the imported PowerPoint data before it is processed by the agent.
- [DYNAMIC_EXECUTION]: The skill employs multiple mechanisms for executing agent-generated code at runtime to manipulate the canvas.
- Sandbox usage:
moda canvas editexecutes local JavaScript files (fix.js) against the live canvas environment (references/edit-code.md). - Markup generation: The
<generate>tag withinmoda canvas markupallows for the execution of JavaScript to emit canvas elements dynamically (references/markup.md). - Constraints: While restricted by character limits and execution timeouts, these mechanisms allow for the execution of logic based on potentially untrusted input.
Audit Metadata