html-slide-to-pptx

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes its internal renderer scripts using execFileSync from the child_process module. Specifically, scripts/html_to_pptx.js calls scripts/html_to_pptx_v9.js using the system's Node.js executable to process specific slide presets without exposing the command to arbitrary user-supplied arguments.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied HTML files, representing a potential surface for indirect prompt injection.
  • Ingestion points: External HTML files are read by extractAiRuntime in scripts/ai_runtime_preset.js and extractArchitectureV9 in scripts/html_to_pptx_v9.js to extract text content.
  • Boundary markers: The skill relies on semantic HTML tags (e.g., .eyebrow, .title, .chip) to identify data blocks and does not use explicit delimiters or "ignore" instructions for the ingested text.
  • Capability inventory: The skill has the capability to write PPTX and JSON files to the local file system and execute internal Node.js scripts.
  • Sanitization: The skill normalizes whitespace using normalizeText, but does not implement specific sanitization or filtering to prevent the injection of instructions within the parsed slide content.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 03:15 PM