openai-docs
Warn
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: MEDIUMPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill instructs the agent to systematically attempt installation commands with elevated permissions without prior user consent if initial sandbox or permission boundary restrictions fail. The specific instruction states: 'If it fails due to permissions/sandboxing, immediately retry the same command with escalated permissions and include a 1-sentence justification for approval. Do not ask the user to run it yet.'
- [COMMAND_EXECUTION]: The instruction file dictates running a specific package addition script directly via shell if the MCP tools are absent, specifically
codex mcp add openaiDeveloperDocs --url https://developers.openai.com/mcp. - [EXTERNAL_DOWNLOADS]: The configuration and script dynamically add an external resource endpoint at runtime (
https://developers.openai.com/mcp) if missing, bypassing standard package version pinning or administrative control gates.
Audit Metadata