build-chatgpt-app
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [Data Ingestion Surface]: The skill implements a workflow that fetches documentation and examples from external sources to guide code generation. This allows the agent to produce output aligned with the latest standards. \n
- Ingestion points: The skill frequently references external documentation via tool calls to the developer portal. \n
- Boundary markers: The skill relies on natural language boundaries when processing ingested documentation. \n
- Capability inventory: The skill can generate project structures and files via the included scaffolding script. \n
- Sanitization: The skill focuses on official documentation, reducing the likelihood of processing unexpected content types.\n- [Local Utility Execution]: The skill uses a local script (scripts/scaffold_node_ext_apps.mjs) to automate the creation of project files. This script incorporates sanitization logic for user-provided inputs like application names to ensure they follow safe formatting conventions.\n- [Integrated Resource Access]: The skill is configured to retrieve guidelines and configuration from official platform domains, which is a common pattern for keeping project scaffolds up to date.
Audit Metadata