session-logs

To search past conversation history, this Skill queries JSONL session logs using jq and rg to filter text, roles, timestamps, and costs, extracting matching messages from stored transcripts. It locates specific phrases across one or multiple sessions in seconds.

Finding context from a parent conversation involves searching stored JSONL session logs to recover earlier chat context. The Skill filters historical references by keywords, retrieves relevant user and assistant turns, and returns the recovered context for the current chat.

Analyzing session logs to measure message counts and summarize costs is supported by querying JSONL transcripts. The Skill lists sessions by date, inspects assistant or user turns, and extracts timestamp and cost data using jq and rg.

Yes, jq and rg are required dependencies to search JSONL transcripts. The Skill relies on jq to parse JSONL files under the OpenClaw sessions directory and rg to perform cross-session keyword lookup and filter relevant text fields.

The best way to perform cross-session keyword lookup in JSONL files is using this Skill, which applies rg for rapid text search across all stored JSONL transcripts and jq to filter the matching results by role, timestamp, or cost.

A limitation of using rg and jq for session log analysis is that queries are restricted to text, roles, timestamps, and costs within the JSONL files under the OpenClaw sessions directory, meaning data outside these stored transcripts cannot be searched.