tech-writer

Warn

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill requires the agent to generate and execute Python scripts to create themed SVG banners for documents.
  • [COMMAND_EXECUTION]: It uses the Mermaid CLI (mmdc) to render diagrams from text descriptions.
  • [DATA_EXFILTRATION]: Locally generated assets are uploaded to Google Drive using gws-drive-upload for document integration.
  • [PROMPT_INJECTION]: The skill processes unvalidated user input (Business Context) which creates an indirect prompt injection surface that could influence generated scripts or document content. Evidence: 1. Ingestion points: Context Gathering phase (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: gws-docs, gws-slides, mmdc, gws-drive-upload, and Python script generation/execution. 4. Sanitization: No sanitization or input validation process is defined.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 10, 2026, 10:17 PM