session-logs

To search past conversation history, you can parse JSONL session transcripts using the `jq` and `rg` command-line tools to extract and retrieve specific historical context, user messages, and assistant responses.

Yes, you can analyze token usage and cost data from agent session transcripts by applying `jq` queries to parse the structured JSONL log files and calculate the associated expenses.

Yes, you need to install both `jq` and `rg` because the skill requires these command-line tools to parse structured log data and execute pattern matching on the conversation JSONL files.

The best way to retrieve historical context from older parent conversations is to search the stored JSONL session transcripts with `rg` for pattern matching and `jq` to extract the relevant message payloads.

Searching session transcripts with `jq` and `rg` works by using `rg` to rapidly locate keyword matches across raw JSONL files and `jq` to parse the matching JSON objects into readable message content.

A limitation of analyzing conversation history from JSONL files is that you are restricted to querying static transcript data, meaning you cannot interact with the live session or modify the historical agent responses.