A specialized utility for querying, auditing, and extracting insights from your AI agent's complete conversation history using JSONL analysis.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install session-logs
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
Copy this prompt to OpenClaw to install it automatically.
Help me install session-logs using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
The session-logs skill provides a robust interface for searching and analyzing historical conversation data stored in session JSONL files. It enables users to bridge the gap between current interactions and past context by providing programmatic access to every exchange, metadata, and tool usage record generated by the agent.
By leveraging powerful CLI tools like jq and rg, this entry in the Openclaw Skills ecosystem transforms raw log files into actionable intelligence. Whether you need to recover a specific piece of information from a weeks-old thread or audit the financial overhead of your automation, this skill provides the precision needed for professional agent management.
To use this skill, ensure that your environment has jq and ripgrep (rg) installed. These binaries are essential for the Openclaw Skills log processing capabilities.
# Install dependencies on macOS
brew install jq ripgrep
# Install dependencies on Ubuntu/Debian
sudo apt-get install jq ripgrep
No further configuration is required as the skill automatically detects the session path at ~/.clawdbot/agents/<agentId>/sessions/.
The skill interacts with JSONL files where each line represents a discrete event or message. The schema is optimized for high-speed filtering.
| Field | Type | Description |
|---|---|---|
type |
string | Defines the record type: "session" (metadata) or "message". |
timestamp |
string | ISO-8601 formatted timestamp of the event. |
message.role |
string | Identifies the sender: "user", "assistant", or "toolResult". |
message.content |
array | Contains text blocks, thinking blocks, or tool call objects. |
message.usage.cost.total |
number | The calculated USD cost of the specific response. |
Loading
An automated research engine that empowers AI agents to explore external data sources, analyze technical trends, and generate structured development roadmaps.

An automated tool for querying restaurant data, ratings, and pricing from Dianping via Openclaw Skills.

A sophisticated memory management and consciousness exploration framework designed to provide AI agents with structured self-awareness and proactive growth capabilities.

Rapidly access concise, community-driven command-line manuals to streamline your development workflow.

An automated discovery and analysis utility designed to track emerging GitHub repositories and technological shifts in real-time.

A specialized monitoring tool for 36kr that tracks trending Chinese tech news, startup financing, and deep industry-specific insights.








































