scribe
Autonomous session scribe — reads today's OpenClaw session logs, extracts decisions, preferences, framework sentences, and project updates, then writes a structured daily memory file. Use when setting up automated memory extraction for an AI agent, or when manually triggering a memory consolidation pass. Works as a cron job (isolated session) or on-demand.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/billc8128/openclaw-scribeWhat This Skill Does
The Scribe skill acts as an autonomous stenographer for your OpenClaw environment. Its primary function is to monitor your daily session history, filter out the technical noise of system heartbeats, and distill meaningful information into human-readable, persistent markdown memory files. By tracking decisions, preferences, framework sentences, project updates, and pending todos, Scribe ensures that your agent retains context across long-term development cycles without needing to rely on transient context window buffers. It transforms raw, chaotic session logs into a clean, structured repository of knowledge.
Installation
To begin, copy the scribe skill files into your workspace using: cp -r skills/public/scribe ~/.openclaw/workspace/skills/public/scribe. Once located, you can automate the process by running the setup script python3 skills/public/scribe/scripts/setup-cron.py. This command registers a nightly cron job that triggers at 23:30, ensuring your memory files are generated automatically after you conclude your work. For ad-hoc memory consolidation, you can run python3 skills/public/scribe/scripts/scribe.py at any point during your session. Configuration can be customized via environment variables, allowing you to change the model, the lookback duration, or enable appending to a global long-term memory file.
Use Cases
Scribe is ideal for users working on complex, multi-day development projects where maintaining consistent state is critical. It is perfect for developers who often lose track of architectural decisions made in previous sessions, or for users who want a structured log of their prompt engineering iterations. By using Scribe, you turn OpenClaw from a single-session tool into a persistent, evolving agent that actually remembers the rules and preferences you have set for it over time.
Example Prompts
- "Scribe, look back at the session from yesterday and ensure all technical constraints we discussed regarding the API integration are captured in the memory file."
- "Run the scribe script now; I just finished a long brainstorming session and want to make sure the project requirements are locked in for tomorrow."
- "Scribe, consolidate today's session, but make sure to focus specifically on the framework rules we defined for the frontend styling."
Tips & Limitations
The most important aspect of Scribe is its configuration. While the defaults work perfectly for most, ensure that your OPENROUTER_API_KEY is correctly set in your environment, as the skill relies on this to classify and extract information from your logs. Keep in mind that Scribe filters out noise, but it relies on an LLM to distinguish between a casual comment and a 'framework sentence'. If you find it is missing critical information, you may need to adjust the SCRIBE_MODEL to a more capable reasoning engine. Additionally, remember that these memory files are plain text; they don't automatically update the agent's active system prompt, so you should review them periodically to ensure alignment.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-billc8128-openclaw-scribe": {
"enabled": true,
"auto_update": true
}
}
}Tags(AI)
Flags: file-write, file-read, external-api