feat(memory): add interactive TUI browser for exploring memory content

[IanKonlog]

Add a new 'agentcore memory browse' command that provides an interactive terminal UI for exploring AgentCore Memory content.

Features:

• Navigate through actors, sessions, events (STM) and namespaces, records (LTM)
• Keyboard navigation: ↑↓ navigate, Enter select, b back, h home, v verbose, m more, q quit
• Glass UI design with Rich tables and cyan highlighting
• Proper API pagination for all list operations (actors, sessions, events, records)
• Breadcrumb display showing current navigation path
• Cursor preservation when navigating back

Also includes:

• Pagination support for list_actors and list_sessions APIs
• Tests to maintain 88% coverage threshold

Description

Brief description of changes

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Performance improvement
• Code refactoring

Testing

• Unit tests pass locally
• Integration tests pass (if applicable)
• Test coverage remains above 80%
• Manual testing completed

Checklist

• My code follows the project's style guidelines (ruff/pre-commit)
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• Any dependent changes have been merged and published

Security Checklist

• No hardcoded secrets or credentials
• No new security warnings from bandit
• Dependencies are from trusted sources
• No sensitive data logged

Breaking Changes

List any breaking changes and migration instructions:

N/A

Additional Notes

Screen.Recording.2026-02-11.at.5.04.39.PM.mov

[theumbrella1]

nit: add test coverage on actual TUI interactions with keyboard bindings (do we have integ tests to update?)

[theumbrella1]

When looking at the memory events, I wonder if we can combine the conversational content in one entry, so that user doesn't have to go into each individual entry and view only one event at a time? (events would be grouped by actor id and session id, and all the events would be consolidated into one json)

[jariy17]

I'm going to merge this PR. This creates list events objects like so

[
    {
      "eventId": "0000001770674253605#92ad3147",
      "eventTimestamp": "2026-02-09T16:57:33",
      "payload": [
        {"conversational": {"role": "ASSISTANT", "content": {"text": "Three."}}},
        {"conversational": {"role": "USER", "content": {"text": "Say four."}}},
        {"conversational": {"role": "ASSISTANT", "content": {"text": "Four."}}}
      ]
    },
    {
      "eventId": "0000001770674246605#f110dd58",
      "eventTimestamp": "2026-02-09T16:57:26",
      "payload": [
        {"conversational": {"role": "USER", "content": {"text": "Say one."}}},
        {"conversational": {"role": "ASSISTANT", "content": {"text": "One."}}},
        {"conversational": {"role": "USER", "content": {"text": "Say two."}}},
        {"conversational": {"role": "ASSISTANT", "content": {"text": "Two."}}},
        {"conversational": {"role": "USER", "content": {"text": "Say three."}}}
      ]
    }
  ]

Can this PR handle it?

[IanKonlog]

When looking at the memory events, I wonder if we can combine the conversational content in one entry, so that user doesn't have to go into each individual entry and view only one event at a time? (events would be grouped by actor id and session id, and all the events would be consolidated into one json)

I am assuming this is the change referenced

[
    {
      "eventId": "0000001770674253605#92ad3147",
      "eventTimestamp": "2026-02-09T16:57:33",
      "payload": [
        {"conversational": {"role": "ASSISTANT", "content": {"text": "Three."}}},
        {"conversational": {"role": "USER", "content": {"text": "Say four."}}},
        {"conversational": {"role": "ASSISTANT", "content": {"text": "Four."}}}
      ]
    },
    {
      "eventId": "0000001770674246605#f110dd58",
      "eventTimestamp": "2026-02-09T16:57:26",
      "payload": [
        {"conversational": {"role": "USER", "content": {"text": "Say one."}}},
        {"conversational": {"role": "ASSISTANT", "content": {"text": "One."}}},
        {"conversational": {"role": "USER", "content": {"text": "Say two."}}},
        {"conversational": {"role": "ASSISTANT", "content": {"text": "Two."}}},
        {"conversational": {"role": "USER", "content": {"text": "Say three."}}}
      ]
    }
  ]

It can be incorporated

[jariy17]

UX: Unclear error message in _resolve_memory_config() (commands.py:102-108)

The error hint:

2. --agent AGENT_NAME (requires config with memory)

should clarify that when --agent is not provided, it defaults to the default_agent set in .bedrock_agentcore.yaml. Something like:

2. --agent AGENT_NAME (defaults to default_agent in .bedrock_agentcore.yaml)

Also, the error doesn't distinguish between "no YAML found" vs "YAML found but memory_id is not set." If the user is in a directory with a valid config but hasn't run agentcore launch yet, this message is confusing — it tells them to do what they already did.

[jariy17]

UX: Home page navigation is confusing (browser.py:234-246)

The starting page shows:

[a] Actors (STM)
[n] Namespaces (LTM)

This isn't discoverable — you have to know to press a or n. Meanwhile the bottom control bar shows [↑↓] navigate [Enter] select, but there's nothing to navigate or select on this screen. The up/down/enter keys do nothing on the home view.

These should be selectable list items instead, so the user can arrow-key to "Actors (STM)" or "Namespaces (LTM)" and press Enter — consistent with how every other view in the browser works.