chatgpt-apps

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill is designed to fetch current documentation and code examples from official OpenAI infrastructure, including developers.openai.com and the github.com/openai organization. These are recognized as trusted sources and well-known services.
  • [DYNAMIC_EXECUTION]: The skill includes a local utility, scripts/scaffold_node_ext_apps.mjs, which generates a project directory with boilerplate code (Node.js server, TypeScript configuration, and a vanilla HTML widget). The script performs input sanitization (slugification) on user-provided arguments and uses static templates for file generation.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external content from documentation and repositories to guide code generation, representing a potential surface for indirect injection.
  • Ingestion points: External documentation fetched via the openaiDeveloperDocs tool and third-party code examples referenced in references/upstream-example-workflow.md.
  • Boundary markers: The workflow encourages citing documentation URLs, though no programmatic delimiters are defined for the ingested text.
  • Capability inventory: The skill can generate project files and write to the local filesystem using the included scaffold script.
  • Sanitization: The scaffold script normalizes tool and application names to prevent injection during file creation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 06:54 AM
Security Audit — agent-trust-hub — chatgpt-apps