feat(core): allow separate orchestrator and worker model config

[Mizelan]

Summary

• add orchestratorAgentConfig to project config schema/types
• use orchestratorAgentConfig for orchestrator sessions while keeping worker sessions on agentConfig
• keep fallback behavior: if orchestrator config is missing, use agentConfig
• add tests for spawn/restore behavior and config validation
• document the new key in config examples

Verification

• pnpm --filter @composio/ao-core test -- src/__tests__/config-validation.test.ts src/__tests__/session-manager.test.ts
• pnpm build

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0e0b79cd10

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Merge orchestrator overrides with base agent config

orchestratorAgentConfig is treated as a full replacement, so any keys omitted there are dropped instead of inheriting from agentConfig. In practice, if a project keeps shared settings (for example permissions: skip) in agentConfig and sets only an orchestrator model override, orchestrator spawn/restore will silently lose those shared settings and launch with different behavior than worker sessions; this contradicts the documented “override (falls back to agentConfig)” expectation and can change runtime safety/approval semantics.

Useful? React with 👍 / 👎.