hermes - 💡(How to fix) Fix Showcase: I built a "Task-Centric Memory" Skill (Auto-categorization + Tiered Compression) [1 participants]

TL;DR

The "Task-Centric Memory" skill may help alleviate the issue of Hermes being slow and forgetful during long conversations by providing features like auto-categorization, tiered compression, and instant search.

Guidance

• To potentially resolve the issue, try installing the "Task-Centric Memory" skill from the provided open-source repository.
• Verify the skill's effectiveness by testing it with long conversations and observing if the agent's performance improves.
• Ensure the skill is properly installed by dropping it into the ~/.hermes/skills/ folder as instructed.
• Test the skill's features, such as auto-categorization and instant search, to understand their impact on the agent's memory and performance.

Example

No code snippet is provided as the issue does not require a code-level fix, but rather the installation and testing of a custom skill.

Notes

The effectiveness of the "Task-Centric Memory" skill in resolving the issue may vary depending on the specific use case and conversation scenarios. It's essential to test the skill thoroughly to determine its impact.

Recommendation

Apply the workaround by installing and testing the "Task-Centric Memory" skill, as it is a community-provided solution specifically designed to address the "messy memory" issue in Hermes.