engram-evomap

What This Skill Does

The engram-evomap skill is the designated long-term memory hub for OpenClaw agents. By implementing the AEIF (Agent Experience Interchange Format) v1.0 protocol, it creates a persistent knowledge repository that allows agents to learn from past errors and successful task resolutions. Unlike temporary session memory, engram-evomap acts as a "genetic" archive for an agent's problem-solving history. When a system error occurs, the skill automatically performs a semantic search through historical capsules to retrieve relevant patches or workarounds, preventing the repetitive "re-solving" of known issues. It effectively turns your agent into an evolving system that grows more capable with every completed task.

Installation

To integrate this skill into your environment, use the OpenClaw repository manager. Run the following command in your terminal:

clawhub install openclaw/skills/skills/404-unknow/engram-2

Ensure that your OpenClaw runtime has appropriate write permissions to the memory storage path, as the skill requires access to disk to persist AEIF capsules.

Use Cases

• Continuous Debugging: Developers can use engram-evomap to store fixes for intermittent build errors or environment-specific configuration hurdles.
• Complex Task Optimization: Automate the recall of specific multi-step workflows that frequently require fine-tuning.
• Team Knowledge Sharing: Since the AEIF format is standardized, memory capsules can be shared or migrated across different agent instances to maintain a shared "tribal knowledge" base.
• Error Interception: Automatically mitigate common runtime failures like SSL timeouts or 404 access errors by retrieving known remediation steps immediately upon failure detection.

Example Prompts

1. "!exp consult 'I am getting a persistent EACCES error when trying to write to the data directory during the build process.'"
2. "I just finished setting up the new deployment pipeline. Please run !exp commit so we can save this workflow for future sessions."
3. "!exp list. Show me the recent memory capsules and tell me which ones have low trust scores so I can clean them up."

Tips & Limitations

• Proactive Committing: The !exp commit command is most effective when executed immediately after a successful, complex multi-step task to ensure the context is fresh.
• Feedback Loop: Utilize !exp score <id> --bad to prune low-quality or outdated advice. Over time, this keeps the retrieval engine clean and increases the signal-to-noise ratio in your error interceptions.
• Privacy: Be aware that memory capsules may contain configuration snippets. Ensure that you do not commit sensitive credentials or API keys that should remain in your .env files rather than the AEIF memory store.