Audit enterprise deployment docs for Anthropic/Claude assumptions and add Codex/OpenAI guidance

[ethanolivertroy]

Context

docs/ENTERPRISE-DEPLOYMENT.md and related README links reference Claude Platform, Anthropic Trust Center, AWS Bedrock, Vertex AI, and Claude Cowork. Some of this may remain relevant historically, but this fork needs Codex/OpenAI-oriented guidance and a clear statement of what is no longer applicable.

Tasks

• Audit docs/ENTERPRISE-DEPLOYMENT.md, README, and related docs for Anthropic/Claude-specific guidance.
• Add Codex/OpenAI deployment and enterprise usage notes where accurate.
• Mark Claude-only deployment options as source-fork/legacy or remove from the Codex path.
• Add a capability-differences section covering model/harness assumptions, tool access, repository context, and terminal behavior.
• Include compliance note: do not imply equivalent FedRAMP/HIPAA/SOC 2 posture without verified source documentation.

Acceptance Criteria

• Enterprise docs no longer imply Codex users should configure Claude Code, Bedrock Claude, or Anthropic Vertex unless intentionally using the Claude source path.
• OpenAI/Codex claims are sourced or scoped conservatively.