fix: correct ability executors creation

[rubinkrief]

Description

The create_linux_ability and create_windows_ability functions in mcp_server.py produce abilities with swapped name and platform values in the executor object. This causes the Caldera agent to silently skip all abilities created through the MCP plugin, since the agent cannot match the executor to its platform.

These fixes swapped between the name and the platform values in the executor object created in both aforementioned functions and now the abilities can be executed in the operation run.

Closes #10

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

1. Started Caldera with the MCP plugin enabled
2. At least one active Linux sandcat agent (or a Windows agent for the Windows test) (Tested on GNS3 tolpology)
3. Used the MCP operation planner plugin to create multiple Linux abilities covering different tactics (e.g., discovery, collection, credential-access), using them to create an adversary and operation.
4. Confirm in the Caldera UI under Adversaries in the generated adversary name that each ability include executors in the correct format.
5. Confirm in the Caldera UI under Operations in the generated operation that:
   • The operation state progresses from "running" to "finished"
   • Each ability transitions through queued → running → finished (not skipped)
   • Each ability produces output visible in the operation link results

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works

[github-actions]

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 5 days