LLM agents don't need bigger memory. They need a cognitive memory kernel.
-
Updated
Sep 8, 2026 - Python
#
neurostate
Here are 8 public repositories matching this topic...
Cognitive security simulation: subliminal carrier + 3-persona GDC debate + chaining evaluator. Educational/research framework.
-
Updated
May 25, 2026 - Python
External observability for synthetic cognitive layers around LLM agents.
-
Updated
May 22, 2026 - Python
A small meta-cognitive simulator for choosing actions by forecasting future synthetic state.
-
Updated
May 22, 2026 - Python
Semantic task recognition for LLM agents — déjà vu fires when a new task matches a past success 🌀
-
Updated
May 25, 2026 - Python
REM-sleep log compression for LLM agents — compress raw logs into lessons, purge the rest 💤
-
Updated
May 25, 2026 - Python
A subjective time perception layer for LLM agents — time flies in flow, crawls under stress.
-
Updated
May 25, 2026 - Python
Rewrite failed memories as legendary successes. The agent remembers a history that never happened — but feels true.
-
Updated
May 25, 2026 - Python
Add this topic to your repo
To associate your repository with the neurostate topic, visit your repo's landing page and select "manage topics."