Prevent clearing of conversation memory after task completion and refine stop handler

[charliecreates]

Overview

This PR addresses two related concerns observed in conversation handling:

1. Refines the conversation stop handler so that stopping execution does not leave the conversation in a terminal state. This ensures the system properly remembers previous user inputs after stopping and resuming a session.
2. Removes all instances of memory clearing or truncation (memory.pop, memory = [], etc.) from task completion handlers, ensuring that past conversation context is never lost, regardless of task transitions.

Details

• Updates the /api/askAI/stop handler to set the conversation state to AWAITNG_USER_INPUT (rather than AWAIT_NEXT_TASK or other terminal states), allowing smooth resumption of conversations without resetting or being blocked.
• Audits all handlers and task-processing logic for any code that modifies, pops, or resets the session.memory array. All such operations are eliminated. Results and inputs are now appended (using push/insert as appropriate) without deleting any prior history.
• Adds a regression test to verify memory retention after stopping, continuing, and completing conversations.

Motivation

Previously, both user-initiated stops and certain task completions would remove memory items (via pop or similar), causing loss of history and breaking continuity for users. This often manifested as the bot being unable to recall previous messages or tasks after a stop or fresh start.

This PR ensures conversation memory is truly persistent throughout a session—neither explicit stops nor task flows can prune earlier messages, guaranteeing accurate and expected recall for the user.

Testing

• Manual testing: Simulate conversation, stop during active dialog, then resume and confirm previous inputs remain accessible.
• Test coverage: New regression tests added for the described scenario(s).

Summary

This change guarantees that session memory is preserved throughout the conversation lifecycle, improving user experience for all flows involving task stops and resumptions.