[codex] Add interruptible sleep tool

[pakrym-oai]

Why

Models sometimes need to pause briefly while waiting for external work, but using a shell command for that delay ties the wait to a process and does not naturally resume when new turn input arrives.

What changed

• add a built-in sleep tool behind the under-development sleep_tool feature
• accept a bounded duration_ms argument, matching the millisecond convention used by unified exec
• end the sleep early when either steered user input or mailbox input arrives
• include elapsed wall-clock time in completed and interrupted outputs
• emit a dedicated core SleepItem through item/started and item/completed
• expose the sleep item as app-server v2 ThreadItem::Sleep and retain it in reconstructed thread history
• regenerate the configuration schema for the new feature flag
• regenerate app-server JSON and TypeScript schema fixtures

Test plan

• just test -p codex-core sleep_tool_follows_feature_gate
• just test -p codex-core any_new_input_interrupts_sleep
• just test -p codex-app-server-protocol
• just test -p codex-app-server sleep_emits_started_and_completed_items

[avinashkamat48]

The implementation validates duration_ms after deserialization, but the tool schema still exposes it as a generic JSON number while the Rust type is u64. That leaves the model/tool caller with a schema that allows fractional values even though parse_arguments will reject them before the explicit range check runs. It would be cleaner to expose this as an integer with min/max constraints if the schema helper supports that, or add a small parse/validation test for fractional input so the model-visible error remains intentional. This matters more here because the tool description says 1..=3600000, but the machine-readable schema currently only says number.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0e0bb90c77

ℹ️ About Codex in GitHub

Codex has been enabled to automatically 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 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[anp-oai]

any harm in having a u64::MAX for the interruptible sleep to avoid the potential for flakes w/ infra weirdness?