para-second-brain

Organize your agent's knowledge base using the PARA method by structuring information into Projects, Areas, Resources, and Archive categories. This logical organization enables full semantic search across local files and ensures persistent, accessible agent memory.

Yes, an agent can search past session transcripts and local notes. The Skill indexes session transcripts and implements a symbolic link trick to enable full semantic search across all stored knowledge, not just immediate memory.

The memory flush protocol prevents context loss during compaction by systematically preserving critical information before memory limits are reached. This ensures specific decisions, research materials, and past conversation details remain accessible.

No, this knowledge management Skill does not require any external dependencies. It operates independently using included scripts and references to structure local files and maintain persistent agent memory without third-party integrations.

Give an AI agent persistent memory across sessions by indexing local files and session transcripts into a structured PARA directory. This approach enables semantic search and prevents context loss, allowing the agent to recall specific decisions made days ago.

An agent loses context and forgets past decisions over time due to limited memory capacity and lack of organized storage. Implementing a memory flush protocol and indexing past conversations into a searchable knowledge base solves this continuity problem.