woostack-visualize

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill reads local files and directories to generate visualizations. It mitigates exfiltration risks by enforcing a strict 'offline-only' constraint in SKILL.md, requiring all styles, diagrams (SVG/CSS), and JavaScript to be inlined with no external network fetches allowed in the generated HTML.
  • [COMMAND_EXECUTION]: The skill performs file system operations including reading source files and writing HTML renders to the .woostack/visuals/ directory. It minimizes risk by requiring explicit user consent before opening the generated file in a browser.
  • [PROMPT_INJECTION]: The skill processes untrusted content from user-specified sources, creating an indirect injection surface.
  • Ingestion points: The parameter (file, glob, or directory) is read and processed in SKILL.md.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified for the source content.
  • Capability inventory: The skill can read arbitrary files and write HTML files to the disk.
  • Sanitization: The impact of potential injection is mitigated by the mandatory offline render constraint, which prevents injected scripts from fetching remote payloads or exfiltrating data via the network.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 08:29 PM
Security Audit — agent-trust-hub — woostack-visualize