feat: synchronize PTY terminal interactions

[konard]

Closes #174.

Reproduction

The PTY capture layer added in #176 could send raw input and resize a terminal,
but interaction readiness accepted literal substrings only and applied the next
key immediately. Agentic TUIs repaint asynchronously, so tests could race a
partial render and had no built-in regex or quiescence synchronization.

The new JavaScript regression starts a raw-mode TTY, emits a delayed readiness
marker, and verifies that a raw Down-arrow sequence is not sent until the regex
matches and output has remained idle. The Rust terminal test exercises the
equivalent regex and idle controls.

Changes

• accept JavaScript RegExp readiness markers and Rust after_regex patterns
• add per-interaction idleMilliseconds / idle_duration quiescence waits that
  restart whenever PTY output arrives
• preserve raw key, text, resize, TERM, live trace, and programmatic capture
  behavior in both implementations
• document the complete PTY testing API and add a runnable raw-mode menu example
  that navigates to web_search and asserts on captured output
• add JavaScript and Rust release fragments

Verification

• PATH=<jq-bin> bun run test — 783 passed, 5 skipped, 0 failed
• bun run check
• bun js/examples/tui-e2e.mjs
• cargo test --manifest-path rust/Cargo.toml --all-features
• cargo clippy --manifest-path rust/Cargo.toml --all-targets --all-features
• cargo fmt --manifest-path rust/Cargo.toml --all -- --check
• cargo package --manifest-path rust/Cargo.toml --allow-dirty
• BASE_REF=main bash .github/scripts/check-language-parity.sh

[konard]

Working session summary

Implemented and finalized PR #182.

Key changes:

• Regex-based TUI readiness matching.
• Output-quiescence waits before interactions.
• Raw keystroke and escape-sequence regression coverage.
• Matching JavaScript and Rust APIs.
• Runnable web_search menu-navigation example.
• Expanded PTY capture documentation and release fragments.

Verification:

• JavaScript: 783 passed, 5 skipped.
• Rust tests, doctests, clippy, formatting, and package checks passed.
• CI passed across Bun, Node, Linux, macOS, and Windows.
• PR is ready, mergeable, current with main, and the worktree is clean.

---

This summary was automatically extracted from the AI working session output.

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Model: GPT-5.6 Sol
• Provider: OpenAI
• Public pricing estimate: $10.424061

📊 Context and tokens usage:

• 180.5K / 200K (90%) input tokens, 19.2K / 128K (15%) output tokens

Total: (180.5K + 7.8M cached) input tokens, 19.2K output tokens, $10.424061 cost

🤖 Models used:

• Tool: OpenAI Codex
• Requested: gpt-5.6-sol
• Thinking level: low (~7999 tokens)
• Model: GPT 5.6 Sol (gpt-5.6-sol)

📎 Log file uploaded as Gist (2213KB)

• View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.

[konard]

🎉 Auto-merged

This pull request has been automatically merged by hive-mind.

• All CI checks have passed

---

Auto-merged by hive-mind with --auto-merge flag