[Enhancement]: Add JSON output format for 'conversation show' subcommand

[spachava753]

Enhancement Category

CLI commands/flags

Problem or Use Case

When using the conversation show subcommand, the output is currently formatted for human readability. However, users who want to:

• Post-process conversation data with other tools
• Export conversations to different formats
• Analyze conversation patterns programmatically
• Integrate CPE conversation history into other workflows

...cannot easily do so because there's no machine-readable output format.

Proposed Solution

Add a --json or -j flag to the conversation show subcommand that outputs the conversation in JSON format.

Example usage:

# Output conversation in JSON format
cpe conversation show <message_id> --json

# Pipe to jq for processing
cpe conversation show <message_id> --json | jq '.messages[] | select(.role == "assistant")'

# Save to file for external processing
cpe conversation show <message_id> --json > conversation.json

The JSON output should include:

• Message IDs
• Roles (user/assistant)
• Content (text and any tool calls/results)
• Timestamps
• Parent message references (for branching context)
• Model used (if available)

Alternatives Considered

• Manually parsing the current pretty-printed output (fragile, breaks on format changes)
• Directly querying the SQLite .cpeconvo database (requires knowledge of internal schema, not stable API)

Impact Scope

Everyone using CPE

Additional Context

This aligns with common CLI conventions where tools provide --json flags for machine-readable output (e.g., docker, gh, kubectl). It would make CPE more composable in scripted workflows and pipelines.