Lobster Compact for Openclaw

A smart conversation compression tool that preserves deep technical context and project intent when LLM context windows reach their limits.

wangxiaofei860208-source
v1.0.0
Apr 4, 2026
0
732
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install lobster-compact

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install lobster-compact using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Lobster Compact?

Lobster Compact is a specialized utility within the Openclaw Skills ecosystem designed to manage long-form developer conversations. As AI interactions scale, context windows can become cluttered or hit token limits, leading to a loss of critical project details. This skill intervenes by performing a chronological analysis of the exchange, distilling complex technical decisions into a structured summary.

By synthesizing user intent, technical patterns, and feedback, the skill ensures that the AI agent maintains a high-fidelity understanding of the project trajectory. It doesn't just shorten text; it extracts the technical essence of the work, including file names, function signatures, and specific bug fixes, ensuring seamless continuity in long-term development cycles.

Lobster Compact Use Cases

  • Managing long-running coding sessions where the context window is nearing its limit.
  • Transitioning between different phases of a project while keeping critical technical decisions visible.
  • Manually triggering a summary using the /compact command to clear mental overhead.
  • Consolidating multi-turn debugging sessions into a single reference point for the AI agent.

How Lobster Compact Works

  1. Chronological Analysis: The skill scans the entire conversation history to identify core requests, technical concepts, and code patterns.
  2. Context Extraction: It specifically targets high-value data such as file names, function signatures, and previously resolved errors.
  3. Structured Generation: A detailed Markdown summary is created covering primary intents, technology stacks, user feedback, and pending tasks.
  4. Memory Synchronization: The skill integrates with memory tools to update MEMORY.md and daily logs, preventing information loss across sessions.
  5. Validation: The process concludes once the summary captures all critical context and the long-term memory files are successfully updated.

Lobster Compact Setup

To integrate this skill into your workflow, ensure your environment supports Openclaw Skills and install the compact package via your CLI:

# Install the lobster-compact skill
openclaw install lobster-compact

# Manual trigger in conversation
/compact

Lobster Compact Data Schema & Taxonomy

The skill organizes summarized data into a highly structured taxonomy within the conversation and external memory files:

Section Data Captured
Technical Concepts Frameworks, libraries, and architectural tools used.
Files & Code Specific file paths and critical code snippets or function signatures.
Fixes & Feedback Errors encountered, their solutions, and explicit user preferences.
Task Pipeline To-do lists, current active focus, and immediate next steps.
Memory Storage Updates to MEMORY.md and date-stamped logs (e.g., memory/YYYY-MM-DD.md).

Lobster Compact Advanced Features

  • Auto-Trigger Mechanism: Automatically activates when the AI agent detects the context window is approaching its technical limit.
  • Focused Compaction: Accepts optional arguments to focus the summary on a specific module or technical area.
  • Deduplication Logic: Utilizes memory_search to ensure that summarized information does not create redundant entries in long-term storage.
  • Multi-Lingual Support: Capable of processing and summarizing prompts in both English and Chinese.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*