session-logs

You can search past conversation history by parsing JSONL session logs stored in your local agents directory to locate specific messages, dates, or keywords across multiple sessions. The Skill filters structured log data to retrieve relevant historical context.

The best way to extract specific messages from conversation logs is by applying keyword searches and lightweight analytics directly to JSONL files. This approach surfaces relevant discussions by filtering user and system messages across all stored sessions.

Yes, you can use jq and rg to analyze session logs for auditing. These tools parse and filter JSONL data to locate specific timestamps, message counts, and keywords, allowing you to reconstruct context from past conversations.

To find and summarize old AI conversations, you index session files from your local sessions directory and extract user and system messages. This reconstructs the historical context needed to continue previous discussions.

Yes, this approach works directly with JSONL files stored in the local agent sessions directory. It operates on the file path structure to index and locate sessions without requiring external databases.

A limitation of searching session logs locally is that it relies entirely on the JSONL files currently stored in your agents directory. It cannot retrieve conversations that were deleted or sessions not saved to the local file system.