markdown-html-orchestrator

Pass

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references Google Fonts (fonts.googleapis.com) and Prism.js (cdn.jsdelivr.net, cdnjs.cloudflare.com) as the only permitted external assets for the generated HTML output. These are well-known, reputable services used for styling and client-side functionality without pose risks to the agent environment.
  • [COMMAND_EXECUTION]: All included scripts (doctype_classifier.py, output_path_resolver.py, route_explainer.py) use the Python standard library for path manipulation, regex classification, and configuration management. No arbitrary command execution or shell injection points were found.
  • [DATA_EXFILTRATION]: The skill accesses a local configuration file at ~/.config/markdown-html/design-system.json to manage user preferences and brand tokens. There is no evidence of this data being sent to external or unauthorized endpoints.
  • [PROMPT_INJECTION]: The orchestrator uses a deterministic two-signal routing logic and a structured 'forcing-question' library to confirm user intent. This pattern effectively prevents the agent from being manipulated into unauthorized actions or silent chaining of operations.
  • [SAFE]: The skill follows security best practices by forking context for heavy processing and strictly limiting capabilities based on a line-count threshold (100 lines) and onboarding status.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 14, 2026, 06:50 AM
Security Audit — agent-trust-hub — markdown-html-orchestrator