para-memory

What This Skill Does

The para-memory skill provides OpenClaw agents with a robust, persistent 3-layer memory architecture based on the PARA (Projects, Areas, Resources, Archives) methodology. Because OpenClaw agents do not inherently retain long-term state across independent sessions, this skill enables durable knowledge persistence. By managing a structured file-based knowledge graph (Layer 1), capturing raw session context in daily logs (Layer 2), and extracting high-level insights into a 'tacit knowledge' file (Layer 3), the agent creates a cohesive, evolving personality and operational history. It automates the maintenance of project statuses, user preferences, and institutional knowledge, ensuring that the agent remains context-aware and efficient over time.

Installation

To integrate this memory system into your agent, use the command-line interface provided by OpenClaw:

clawhub install openclaw/skills/skills/aisoftgg/para-memory

Post-installation, ensure your workspace is prepared by executing the setup routines detailed in the skill documentation, which includes initializing the directory structure (life/projects, life/areas, life/resources, life/archives, and memory) and populating the initial templates for tacit knowledge and daily logging.

Use Cases

• Long-term Context Retention: Maintaining consistency in professional relationships by recalling specific user preferences and past project decisions.
• Project Management: Organizing complex workstreams by keeping 'status.md' files updated for every active project within the life/projects directory.
• Agent Onboarding/Self-Correction: Allowing an agent to learn from its own mistakes and evolve its communication style by logging 'lessons learned' into the tacit knowledge layer.
• Knowledge Management: Providing a centralized, searchable repository for research, templates, and reference materials that the agent can access across different work sessions.
• Nightly Consolidation: Automating the transition from raw, noisy daily logs to permanent, structured knowledge during end-of-session cleanup processes.

Example Prompts

1. "Initialize the PARA memory system for our current marketing project and create a new daily note for today's session."
2. "Review my tacit knowledge file and extract any relevant preferences about how I prefer my technical summaries formatted."
3. "Consolidate today's meeting decisions from my daily notes into the status.md for the Alpha Project and update the resource list with the new URL I provided."

Tips & Limitations

• Consistency is Key: The efficacy of this skill depends entirely on the agent's discipline in writing to files. If the agent fails to perform nightly consolidation, the memory system will become cluttered with raw data that is difficult to query.
• File-System Reliance: Since this skill operates via file-write and file-read actions, ensure your agent has appropriate permissions for the directory structure.
• Avoid 'Mental Notes': Train your agent to treat the life/ and memory/ directories as its primary source of truth. Relying on the LLM's context window alone is discouraged for long-term data storage.