fix: cleanup scripts skip DEFAULT endpoint to prevent resource leaks

[jyotsnamas]

Summary

• Fix 18 cleanup.py scripts that silently leak AgentCore runtimes and endpoints
• The API auto-manages DEFAULT endpoints — they cannot be explicitly deleted
• Scripts that attempt to delete DEFAULT get ConflictException, which then prevents runtime deletion
• Scripts report "Cleanup complete" while resources remain alive and accruing costs

Root cause

1. cleanup.py calls DeleteAgentRuntimeEndpoint(endpointName="DEFAULT")
   → ConflictException: "Default endpoints are removed when you delete the agent"
2. cleanup.py calls DeleteAgentRuntime
   → ConflictException: "This agent has endpoints that must be deleted first"
3. Both exceptions caught silently → "Cleanup complete" printed
4. Runtime + endpoint remain READY in AWS

Fix

Skip DEFAULT endpoints during explicit deletion, matching the pattern already used in 01-strands-bedrock/cleanup.py:

for ep in endpoints.get("runtimeEndpoints", []):
    if ep["name"] == "DEFAULT":
        continue
    control.delete_agent_runtime_endpoint(...)

Verification

Deployed and cleaned up multiple modules in us-west-2:

• Without fix: aws bedrock-agentcore-control list-agent-runtimes shows runtime still READY after cleanup
• With fix: Runtime successfully deleted, confirmed gone via list-agent-runtimes

Test plan

• Deploy 02-a2a-protocol → cleanup now deletes runtime successfully
• Deploy 09-mcp-e2e/01-server-e2e → confirmed leaked runtime with old cleanup, deleted with direct API call
• Verified strands-bedrock cleanup (already has the fix) works correctly

🤖 Generated with Claude Code