Add custom request timeout support to @astroid/client fetch wrapper

[Cjay-Cyber-2]

Description — what to implement and the why/impact.

Agents operating in time-sensitive environments must not hang indefinitely if the Astroid API or Stellar network nodes become unresponsive. This issue adds configurable request timeout handling using AbortController to @astroid/client.

Context & Requirements — background, constraints, design references, edge cases; grounded in the actual repo (real files, stack, conventions).

The client should accept a default timeout (e.g. 30000ms) in its initialization options and support per-request timeout overrides. When a request exceeds the timeout, it must abort the underlying fetch operation and throw a clean, typed timeout error.

Acceptance Criteria — a checklist ("- [ ] ...") of specific, testable conditions that define "done"..

• Add timeout option to client configuration and per-request options.
• Use AbortController and setTimeout to abort requests exceeding the configured duration.
• Throw a specific typed error or custom error subclass for timeout failures.
• Add unit tests using fake timers or mock fetch to verify timeout cancellation.

Implementation Guidance — likely files/modules to touch (use real paths from the repo context) and a suggested approach, without over-prescribing.

Review packages/client/src/ for fetch execution logic. Ensure timer handles are properly cleared in finally blocks to prevent event loop leaks.

Testing & Validation — how the contributor should prove it works (tests to add/run, manual checks, screenshots for UI).

Run pnpm test --filter @astroid/client and verify the test suite passes cleanly.

Submission Guidelines — must open a PR that includes "Closes #"; assignment is required before starting; follow the repo's existing style and conventions.

---

Wave complexity: Medium