[MCP] Support request cancellation for in-flight network calls

[manoahLinks]

Summary

Tool handlers issue fetch and Horizon/Soroban calls with no AbortSignal, so a cancelled MCP request keeps running and can still mutate state after the client has given up.

Scope

• Keep the implementation aligned with the existing MindVault architecture.
• Add or update automated tests.
• Update docs when the change affects contributors, deployers, or agent users.

Acceptance criteria

• An abort signal is threaded from the request handler into outbound fetch and RPC calls.
• A cancelled request stops before performing a payment or state write where possible, and tests assert the abort propagates.
• Failure cases are deterministic and safe to show in issue-driven contributor work.
• Relevant local tests pass.

[drips-wave]

@manoahLinks This issue could not be added to the Stellar Wave Program because the per-repo points budget would be exceeded (25000 / 25000 points already used). The label has been automatically removed.

The budget resets at the end of each Wave. You can review the remaining budget for your repos on the Orgs & Repos page.

ℹ️ Learn more about points budgets