Open Second Brain runs alongside Hermes Write Gate out of the box. Your approve/deny on built-in memory stays the source of truth, and our mirror only ever reflects writes you actually committed. Gate it in Hermes, and the second brain follows.
https://t.co/EcgkD9tfRw
Introducing Write Gate in Hermes Agent.
Now you have the capability to be able to approve/deny memory updates, skill updates, and skill creation with the same familiar mechanisms as approving dangerous commands.
If you are using a small model that doesn't always recognize what it learned, a secure environment that needs gating before things that can affect operations occurs, or just want to be more involved in the self improvement process of your Hermes Agent, now you have full control!
This will be included in the next major release version, but you can run `hermes update` now to access early!
v1.3.0 brings Open Second Brain to @opencode
OpenCode can now use the same memory loop I use with other agents: rules in replies, learned context saved, and sessions imported back into the Brain.
OpenCode gets a real long-term memory layer.
https://t.co/8C5kNDJ3aH
@Teknium Not surprising given the incredibly strong Hermes community. I released my stable plugin a few days ago, and Discord members quickly spotted an issue I fixed right away. Onward!
https://t.co/fGnr4PPZEF
Your AI agent's memory shouldn't be a black box. Open Second Brain keeps what it learns from everything you do as plain Markdown in your Obsidian vault - readable, editable, yours. Plugs into Hermes by @NousResearch via the official memory-provider API
⭐️ https://t.co/ae4bIRGlZf
Your AI agent's memory shouldn't be a black box. Open Second Brain keeps what it learns from everything you do as plain Markdown in your Obsidian vault - readable, editable, yours. Plugs into Hermes by @NousResearch via the official memory-provider API
⭐️ https://t.co/ae4bIRGlZf
@Teknium Thanks! I'm already an active member in the Discord. The Open Second Brain plugin even has its own dedicated thread in the forum channel where I regularly post updates and new releases.
@Teknium Merged! 🚀 Open Second Brain is now a 100% first-class memory provider for Hermes. The workaround is gone, the integration is seamless. Shoutout to @NousResearch for the amazing agent framework!
https://t.co/zkCp4UwanD
🚀 Major update for Open Second Brain! The plugin is now ready to use as a first-class memory provider for Hermes.
I built it to give your autonomous agents structured, long-term memory to boost context retention and decision making.
🔗 https://t.co/Z5I5vGAXPE 🧵👇
🛠️ I also opened a PR to fix how memory providers are registered in hermes-agent: https://t.co/prxKl7IGkJ
It currently needs a small workaround. Hoping @Teknium takes a look so we can drop the hack and make the integration fully native!
Just dropped my plugin https://t.co/A5omOgr7JT in the Hermes Agent Discord. Meant to catch the Jam Session live, but I'm on a train. Building on the move - all of life is motion. @Teknium@NousResearch
The most amazing part, although expected, is that this all happens through Dark Factory itself.
It finds signals for new features, generates specs, reviews code, fixes issues, and ships improvements back into the same system.
The loop is becoming very real.
Today Dark Factory shipped 5 OpenSecondBrain releases in a single day.
Not tiny patches. Large releases with 6-11 features each.
A few months ago this sounded impossible to me. Now I’m refreshing the releases page faster than I can fully process what’s happening.
I haven’t made any commits to OpenSecondBrain for several days. I wanted to stop touching it for a bit and quietly test its current behavior.
And honestly, it feels amazing! It works much better than I expected.
So I decided to stop adding new features.
But...
The latest release pushed much further into synthesis and long-term reasoning.
Daily and weekly summaries, long-running change tracking, higher-level synthesis flows, ongoing project evolution awareness.
This is the first time it started feeling like continuous cognition.
If codegraph isn't installed, the agent falls back to the old path and recommends it once, with the real file count from your project. Say "no" and it stops nagging.
The .codegraph folder lands in .gitignore automatically.
In OpenSecondBrain v0.10.13, the agent understands your code even better: it notices an installed codegraph by @ColbySMcHenry and routes structural questions to it - who calls what, where a symbol is defined, what breaks if you change it.
https://t.co/ZXpR7lBvxJ
Previously these questions kicked off a grep + read loop - noisy in tokens, and text search often catches false positives. Now one precise AST query replaces dozens of file reads.